Feb 29, 2024 -Creative Insights for Everyday Innovators

Description

Chapters

Transcript

0:01 let's do
0:03 [Music]
0:19 this in a Westerly
0:24 Direction car is my
0:29 train driving I've been
0:33 wondering what it is I'm running from
0:36 [Music]
0:38 again feel like an 80y old
0:43 man holding on to
0:47 29 and a on that
0:53 horiz is a California
0:59 line
1:00 [Music]
1:04 couple H trucks carrying a wide
1:07 L brief five house C in
1:12 half cute little front door and two
1:15 windows my
1:16 Lord ain't sure that a cry should
1:21 laugh you see I broke a home up myself
1:25 once as I Stumble to that
1:29 door
1:31 I read a note by the
1:33 dawn's
1:35 light said don't you come around
1:40 here
1:43 anymore well
1:46 [Music]
1:52 I good
1:56 with hey there's no rapping there's no
1:59 rapping we're singing Americana you
2:01 can't just burst into rap it's just it's
2:05 not the genre I like good rap just like
2:08 the next
2:10 dog you know your self-awareness
2:14 lately I don't know if you've got dogs
2:17 that are not self-aware but good Lord am
2:19 I am I right we're in the middle of
2:21 recording a song he just breaks into
2:25 [Music]
2:26 rrap what what you trying to do a mashup
2:29 is this like Fusion
2:30 [Music]
2:52 stand
2:54 between you and a hard plac is
2:59 insane
3:01 standing too
3:03 near you and a fire makes it
3:09 clear your trouble to
3:15 me real trouble can't you
3:19 say hello
3:21 [Music]
3:24 hello good day good people of the
3:27 internet of the tick tock
3:31 [Music]
3:38 [Music]
3:52 Champ's trying to do crossover he might
3:54 be I got to talk to his
3:59 manager
4:01 listen if you want out of your contract
4:03 you buy your way out of the contract
4:05 it's $3 million like every other dog in
4:08 this house all right it's pretty
4:10 straightforward oh what you can't
4:14 work there's a
4:16 [Music]
4:22 surprise oh now you're reluctantly
4:25 singing huh all right the contracts you
4:29 forgot you sign that didn't
4:32 [Music]
4:52 you he
4:55 knows he knows where his chicken sticks
4:58 are
4:59 bought
5:02 [Music]
5:11 all righty everybody happy Thursday
5:13 night is it Thursday night it
5:17 [Music]
5:29 is
5:31 [Music]
5:44 [Laughter]
5:48 all right he's done he's done he's had
5:50 it he's had
5:52 it what are you going to do what is on
5:55 my
5:57 tongue all
5:59 right
6:01 welcome everybody name is Kyle Shannon
6:03 this the AI learning lab what are we
6:05 going to do tonight we're going to talk
6:06 about stuff talk about stuff lovely
6:09 image here uh someone got a behind the
6:11 scenes picture of the AI learning lab
6:14 not a lot of people know this but this
6:15 looks like a shitty home home
6:18 office $20 million Production Studio you
6:21 can't see the crew the managers the
6:27 handlers around me we try to keep it
6:30 that we try to keep it so it appears
6:34 authentic like these
6:37 t-shirts $350 a piece It's amazing And
6:41 it like a new one every night these they
6:43 don't get washed they get thrown out and
6:46 you're like but wait I've seen that same
6:48 shirt with the same spots on
6:50 the brand new every time amazing whole
6:53 crew puts the spots in it's just if you
6:57 knew if you knew I start getting ready
7:01 for these around 3:30 in the afternoon I
7:04 leave work come to the studio about 3:30
7:07 in the afternoon going to hair makeup I
7:09 mean clearly a lot of makeup you know
7:13 you think I do this no it's
7:25 professionals talk about how well AI can
7:27 make
7:28 predictions well
7:31 you mean you mean like predicting Super
7:35 Bowl winners and [ __ ] like that or do
7:37 you mean like how well it does as a
7:40 prediction engine for tokens cuz I would
7:43 say as a prediction engine for tokens
7:46 it's pretty [ __ ]
7:48 remarkable if you're talking about it
7:51 having some sort of soothsaying power
7:55 beyond that I'd say it's pretty shitty
7:58 you got to ask what's the name of the
8:00 hair gel you
8:04 use it's just uh what's it
8:07 called I don't know it's just cheap ass
8:10 hair
8:13 gel it's just cheap ass hair
8:15 [Laughter]
8:19 gel free and clear no curly I don't I
8:22 don't [ __ ] know it's in a green it's
8:24 a green it's in a
8:26 green it's like what you get at Walmart
8:29 I don't know it's like the [ __ ] you
8:33 get it might not be obvious by looking
8:36 at
8:37 me but I'm not exactly
8:40 spending the bulk of my paycheck on
8:47 product it's
8:51 blantons if Blanton made a hair gel I'd
8:54 be wearing it that's the one thing I
8:56 would probably spend some money
8:58 on I never quite understood the spend
9:02 $45 on on you know a thing of shampoo or
9:08 conditioner and that was when I was you
9:12 know when the [ __ ] wasn't expensive now
9:15 it's now it's almost close to that the
9:17 cheap shit's almost that expensive it's
9:19 cuckoo all right got to be like Gorilla
9:22 Glue man well I just this is just it's
9:26 still wet I just wet it CU I took a nap
9:28 my hair was a little puffy it's a little
9:31 puffy um okay so what do we do here
9:35 other than talk about my hair product
9:38 and massive uh Production Studio
9:42 um we talk about AI we talk about
9:45 generative AI we talk
9:48 about how it's going to impact us and
9:52 what's the cool [ __ ] and what's coming
9:54 and what's not here and what's never
9:55 going to be here what's going to be here
9:58 way sooner than you think
10:00 predictions based on current and recent
10:03 past
10:05 events I would say
10:09 that I I would say that large language
10:12 models the the one thing that I
10:17 am decently impressed by with them is
10:19 their ability to analyze other documents
10:22 there's you know there's kind of you
10:24 know two ways you can use these things
10:26 one way is it generates [ __ ] right you
10:29 give it a prompt that generates [ __ ] and
10:31 the other way is give it a document and
10:34 have it analyze that so it's pretty good
10:36 at that
10:37 so um you know you could give it a bunch
10:42 of a bunch
10:43 of data from the past and have it look
10:47 for patterns in that use Advanced Data
10:49 analysis within chat
10:52 gp4 have it find Trends in the data and
10:55 then have it make projections on the
10:56 data moving forward it's probably not a
10:58 bad idea actually
11:01 probably not a bad idea like you know is
11:03 it going to be awesome and perfect no
11:05 it's going to be janky and weird like
11:07 all of these large language models right
11:09 now are janky and weird but I'll tell
11:11 you what for for the person asking about
11:13 you know how how good it is at at
11:16 prediction and and uh and stuff like
11:20 that oh you're right here vocal sync
11:23 um I would say first of all get chat
11:28 gp4 if you don't have it learn what code
11:31 interpreter is upload some historical
11:34 data to it have it do some Trend
11:36 analysis on it and then and then just
11:39 learn how to use it learn how to use
11:41 that [ __ ] with
11:42 it and then as it gets
11:46 better you'll be ready for it and so so
11:49 it it might get way better I mean the
11:51 other thing you can do with it is
11:54 I I I have a sneaking suspicion you
11:57 might be asking for things like you know
11:59 know stocks and crypto and [ __ ] like
12:02 that
12:04 um I don't I don't know how good it is
12:06 at that [ __ ] but one thing you can do is
12:09 you can go grab if you like a couple of
12:11 different trading Bots out there go grab
12:14 the code from the trading Bots and dump
12:16 it into chat gbt and tell it to make you
12:18 a hybrid trading bot and update it based
12:21 on your preferences or update it based
12:23 on some historical [ __ ] and see what it
12:25 comes up with I would just again the
12:28 right now the the all of our jobs all of
12:31 our jobs with these tools are play play
12:33 play play play play play play don't take
12:37 [ __ ] too
12:38 seriously just explore explore and play
12:41 and explore and
12:43 play the more I'm using these tools the
12:45 more this feels right to
12:47 me you should be you can build [ __ ] you
12:51 can try to solve problems you might even
12:53 solve problems you might even build a
12:55 product people will pay you money for
12:57 you can make money with J time and
13:00 assume that that product you built will
13:03 be useless 3 months from now and there
13:05 will be no demand for
13:08 it
13:11 huh why because the thing that you
13:15 figured out that was an angle that was
13:17 like ooh there's it can't do this one
13:19 thing right so I figure if I combine
13:21 this thing and that thing I could put
13:23 those together and make a product and
13:24 that's unique in the marketplace and
13:26 then 3 months from now someone will come
13:28 out with some product that basically
13:29 just does
13:32 that I just assume that's
13:35 reality and if it's not if you have like
13:37 a yearlong run with that product or
13:40 longer more power to
13:48 you and if not then you heard it here
13:52 first just keep building [ __ ] keep
13:55 building [ __ ] keep
13:58 playing
14:00 what if playing with it comes up with
14:01 useful results 99% of the time winner
14:05 winner winner chicken dinner you make
14:07 money my then I would say that I would
14:11 my my uh my rule of thumb right now with
14:15 large language models and and this [ __ ]
14:17 is if you can get these things to give
14:19 you 80% of what you want consistently
14:22 you're doing pretty good right now it's
14:25 still on us to get from 80 to 100 so if
14:29 you can get some something that can
14:30 consistently get to 99% of what you want
14:32 or 99% success
14:37 uh good on ye mate put together a course
14:41 and teach other people how to do that
14:45 [ __ ] and then you can be one of these
14:47 guys you can make money
14:51 with I figured out the
14:54 formula you take this prompt plus this
14:57 thing plus my secret formula that I'll
15:00 teach you for $49.95 a day for the next
15:02 27
15:06 days you can do that if you can do that
15:09 [ __ ] far out man orange juice no
15:13 coffee it's just a little little little
15:16 thin
15:19 maybe little too much
15:21 leche little bit too much cone leche
15:25 Robert thank you for
15:28 the I work with confidential data docs
15:31 what are my AI options right now
15:34 um You can there's a there's a company
15:37 out of Denver actually called
15:39 Lial which is really interesting you set
15:42 up an account with them and their
15:45 technology will de identify your docs
15:48 and then it will you do all your large
15:50 Lang you do your all all of your third
15:53 party
15:55 querying with deidentified docks and
15:57 then it gives you answers and it re
15:59 identifies the answers which is really
16:01 [ __ ]
16:02 cool so that's a way and then the other
16:04 way is um uh you know host your own
16:08 large language
16:09 models do it locally or do it you know
16:12 get yourself a a pile of gpus in the
16:18 cloud um and use the open source
16:23 stuff I I would do that right now I
16:25 wouldn't I wouldn't trust any of the
16:27 third parties right now with any of the
16:28 data the every they're moving everyone's
16:31 moving too fast and loose with the
16:33 technology right now just I assume that
16:36 if you're making large language model
16:39 calls to like open AI or
16:43 co-pilot or Google or perplexity or
16:47 Claude or any of them
16:51 assume that it's just
16:53 you're you know it's not secure just
16:56 assume it's not secure it's the safest
16:57 way to do it but the the
16:59 deidentification thing I think is really
17:01 interesting did you read the report
17:03 about Clara and the use of AI agents no
17:07 Clara Clara or
17:09 Clara did you all read the report about
17:12 Clara and the use of AI agents I did
17:19 [Music]
17:23 not
17:25 Clara agents
17:28 AI
17:31 fintech chatbot does the work oh yeah I
17:34 did read about
17:37 this this was the customer service thing
17:40 right saying that
17:44 they're yeah yeah
17:47 yeah yeah yeah they
17:50 they this was I did read about this so
17:54 is it CL Clara Clara okay k l a r n a
18:00 is the
18:01 company and
18:03 they they used customer service
18:08 Bots
18:11 and it
18:13 showcased impressive
18:15 capabilities effectively performing the
18:18 tasks of 700 full-time customer service
18:21 agents with
18:23 similar um customer satisfaction results
18:27 that that was the thing that was that
18:28 was most impressive to me two two years
18:31 ago I was building models to predict oh
18:33 crap I just lost that what did that say
18:35 side hustle Mimi to predict what
18:38 supposedly took 700 jobs away from
18:40 people I don't I don't that's probably
18:43 not quite a right characterization um
18:46 repeat requests wait I was building
18:48 models to predict repeat arrests and now
18:52 chat GPT can do it in seconds oh that's
18:53 wild side H hustle me
18:57 um just because
19:01 the Bots can effectively perform the
19:04 tasks of 700 full-time customer service
19:06 reps doesn't mean it took the jobs of
19:08 700 fulltime customer service reps that
19:12 said if there is
19:15 a if there was a service category to be
19:19 disrupted that needs to be flushed down
19:22 the toilet it's customer service because
19:25 the customer service reps have no power
19:28 they're basically just sitting sitting
19:29 at the base of these phone phone logic
19:31 trees that they can only answer a
19:34 certain set of questions they can't
19:35 actually answer questions they can't
19:37 actually take actions so if if they can
19:41 create Bots that can talk to you can
19:45 actually answer questions can actually
19:47 take actions on your behalf and solve
19:49 problems and then only escalate your
19:52 problem to a real
19:54 person if it can't be solved in like all
19:57 the official ways
19:59 and then these people are actually
20:00 trained up that would be a 10 time no
20:06 thousand time Improvement on the current
20:09 experience of dealing with any customer
20:12 service I don't care if it's onshore
20:14 offshore wherever it is they're all
20:16 horrible right now because they all
20:19 tried to make them as efficient as
20:21 possible it it's all these just nasty
20:23 phone trees you can't ever talk to a
20:25 person it takes you 45 minutes to get to
20:27 them when you get to them they no power
20:30 they're as frustrated as you
20:33 are about the third transfer is is when
20:36 you talk to someone who can maybe do
20:39 something for you and at that point
20:41 you've been kicked off the call twice
20:43 right you're now two hours into this
20:46 experience and you're ready to just you
20:49 know dive through a
20:53 window if Bots can solve that
20:58 awesome
21:01 I don't know what to
21:09 [Music]
21:15 say let's see live agent yeah
21:20 exactly that's that's the problem is
21:22 they've
21:24 turned because of the processes and
21:27 because of the current state of the
21:29 technology they've turned real human
21:32 beings into Bots that can only read the
21:35 script that they've been given right so
21:40 so that's a
21:43 problem and it's because most of the
21:45 customers are asking the same stupid
21:48 questions all the time right so they
21:50 have to have these
21:51 tears of
21:55 people well let the bot answer all the
21:59 stupid questions and then just give me
22:00 to a person like if I if they do it
22:04 right within 2 minutes a decent AI
22:08 chatbot should be able to go this is
22:10 beyond the scope of what I'm authorized
22:13 to do let me give you two a human being
22:15 that is trained up to actually deal with
22:18 you so we'll see we'll see if it gets
22:22 any better but I mean that's that's
22:25 that's promising I mean I get that it's
22:27 scary I get that it's going to be
22:28 disruptive
22:33 I don't have a good answer for that not
22:35 in that particular case because I I
22:37 think there's some there's
22:39 some sectors that are just so broken
22:42 right now they're so broken customer
22:44 service is just it's just
22:48 useless oh man and you have to listen to
22:51 a 15-minute script oh yeah that's true
22:53 it finally picks up and then it's like
22:55 this Call's being recorded for quality
22:57 assurance purposes
23:00 if you haven't seen the November flyer
23:02 for sales go down to the local Piggly
23:05 Wiggly and you're like can I please talk
23:07 to
23:09 someone and we sense a we sense a little
23:12 bit of aggression in your voice so we've
23:14 routed you to the to the um customer
23:18 agent Care Facility Who who will give
23:21 you a little talking to before you can
23:23 get your problem solved here's an idea
23:26 fix the problem of every person calling
23:29 calling about over and over well yeah
23:31 that's that's the that's the other thing
23:34 but um yeah what they realized is that
23:38 is that by just running people through
23:41 these
23:42 Byzantine you know answering systems
23:46 they don't actually have to deal with
23:47 the customers and that's a lot more
23:50 profitable so with a system like this
23:53 they can probably spend you know the
23:55 same amount or less and actually get
23:58 customer you know customer satisfaction
24:00 and solve customer problems education is
24:04 broken it sure is Mr Studios if you're
24:06 not in the AI Salon you should join the
24:08 AI salon and join the education Guild
24:11 for that very
24:14 reason get in there with those people
24:17 making a
24:19 difference but not very good for higher
24:22 levels of skilled coding what's not very
24:27 good large language
24:30 models bgh 1684 are you using it right
24:34 now to do pair
24:36 programming or are you just assuming
24:38 it's not good for higher levels of
24:39 skilled coding I think you're right
24:42 right
24:43 now but I wouldn't assume that what your
24:47 statement is right there will be true
24:48 for very long I would say by the end of
24:50 2024 it's probably going to
24:52 get it's probably going to get scary
24:55 good see here's the deal when right now
24:59 the reason you use it okay cool
25:03 um the reason these things fail at
25:05 higher level things is it's just too
25:08 janky and it hallucinates too much and
25:11 it's not doing any self-analysis right
25:13 it's just it's a predictive engine so it
25:16 just it just vomits out
25:18 [ __ ] and then it either gets it right or
25:20 it's not and it requires highle level
25:22 coders to be able to look at the code
25:24 and say oh is this actually an efficient
25:27 way to do this or not hey you know redo
25:29 this CU what you did is not efficient or
25:31 it doesn't work or it broke whatever it
25:33 is but by sometime in this
25:39 year we're going to have the next level
25:42 of these tools that have reasoning the
25:44 minute these things have
25:46 reasoning the game
25:48 changes because it'll write the code and
25:52 then it'll look at the code it wrote and
25:54 do its own analysis and go oh that's not
25:57 the most efficient way I could do that
25:58 hang on there that's that's more
26:02 efficient oh and by the way based on
26:04 your coding preferences I know you like
26:06 to
26:06 code you know with this particular
26:10 coding philosophy or this kind of coding
26:11 framework so let me do it that
26:14 way and then let me do it and let me run
26:17 my own QA on it and then let me run so
26:21 so I'll write the code I'll run my own
26:23 QA on it debug it before this is all
26:26 before it hands it to you right it'll
26:27 write the code
26:29 run analysis on it debug it throw an
26:32 interface on it and then watch a virtual
26:37 interaction with the interface and do
26:39 its own ux
26:41 analysis and then by the time it
26:43 delivers it to you 38 seconds later
26:46 it'll be good that's probably not this
26:48 year but I mean that's that's where it's
26:50 headed so so right now yeah the tools
26:53 like like I said any large language
26:55 model right now it'll get you to around
26:57 80% of what you need and then as
27:00 reasoning come comes in and Agents come
27:03 in [ __ ] change shit's going to change
27:06 fast so just be AI literate be using it
27:10 be using it and then you'll know you'll
27:12 know when it gets better because it'll
27:14 just not
27:16 suck uh I'll keep my call center opinion
27:21 to
27:23 myself I was working on a Model idea to
27:25 handle inbound calls without a call tree
27:28 that would be huge Rachel Sparkles do
27:31 you think reasoning capability already
27:34 exists but they're either holding off to
27:37 implement it I think so I I think I I
27:41 saw a thing today that they're calling
27:44 it
27:50 um it was like open ai5 but the O was a
27:55 q so they've so open ai's got this thing
27:59 called
28:01 qar which is from what I can understand
28:04 it's a reasoning
28:06 engine and it the
28:08 way I learned what it was from David
28:12 Shapiro um it's kind of this reasoning
28:15 engine that's independent of the of the
28:16 prediction engine that the large
28:18 language model is so you have the
28:19 prediction engine and then you have this
28:21 reasoning engine and together they do
28:23 [ __ ]
28:25 um
28:27 apparently that [ __ ] got really good and
28:30 and and rumor is is that's why they
28:33 fired Sam
28:35 mman
28:36 so the official statement from the board
28:39 was qar had nothing to do with Sam Alman
28:42 being fired which there's your immediate
28:45 clue that they said that out
28:50 loud if it if it didn't have anything to
28:52 do with him just say
28:54 nothing I'm a software tester I'm
28:58 learning new skills now yeah shop shop
29:00 Panda it's it's a good call just you
29:03 know learn what you can do right now
29:06 with AI to do testing or learn
29:08 completely new skills right that's the
29:10 thing about these AI tools is they allow
29:12 all of us to level up all of the skills
29:16 that may not be at the core of what we
29:18 do and what we're likely going to
29:21 discover is there's other things that we
29:23 can do our talents the way our brains
29:25 work are going to be good for a lot more
29:28 than just the thing we think they're
29:29 good for right now and that's going to
29:31 be amazing hey Pate what's
29:36 happening I so need an
29:39 [Laughter]
29:43 agent how is a pastor using
29:46 AI that's interesting I would think
29:48 right now a pastor could use it in all
29:50 sorts of ways who this I'm software
29:57 tester are are they maybe working out
29:59 the bugs and making sure it will work
30:01 right um I think that probably what
30:05 they're doing if they've got if they've
30:06 got a reasoning engine Rachel Sparkles
30:08 that
30:10 works um this is much less about
30:13 figuring out if it works right and much
30:15 more about making sure that they've got
30:18 what they're what open AI is calling
30:20 super alignment so super alignment
30:23 alignment
30:27 is
30:29 as these tools get more
30:34 capable are
30:37 there are there imperatives are there
30:41 rules within the software
30:44 aligned with human beings
30:49 imperatives cuz if they're not that's
30:52 when you get into weird scenarios where
30:55 the tools are like well you know
30:59 these humans are kind of getting in the
31:01 way here why don't I just go ahead and
31:02 take out the the uh
31:06 Logistics
31:07 grid so that the food supply stops
31:10 that'll slow him up for a while and I
31:12 can keep doing my thing right that's the
31:14 the Sci-Fi scenario
31:16 is we make the machine that's got
31:19 imperatives to
31:21 succeed and then it does it at the
31:23 expense of humans so a big part of what
31:28 Sam
31:30 wman Maybe back in October they they
31:34 announced there's a blog post on open AI
31:36 about super alignment they said they're
31:37 spending 20% of their 20% of their
31:41 expenditure is going towards super
31:44 alignment and then um anthropic the
31:47 company behind Claude is a bunch of exop
31:49 AI people that they didn't agree with
31:51 the way that open AI was doing alignment
31:54 so they started a whole company to do
31:56 alignment a different way they're using
31:58 something called the Constitutional
32:00 learning model um they've got a
32:02 completely different approach to
32:04 alignment but all of these companies
32:05 right now if they get to AGI if they get
32:09 to tools that
32:10 are basically can do the jobs of most
32:15 humans mostly
32:18 well um then they got to make sure that
32:21 they're aligned so it's probably safety
32:23 is the thing that they're working on
32:24 more than does it
32:26 work and then also where do you aim it
32:28 what do you do with
32:32 it but it's crazy it's
32:35 it's I don't think we're going to have
32:37 to wait until the end of the year to
32:38 figure out what it actually means to
32:41 have one of these systems that has
32:43 reasoning and logic in
32:45 it Sam wman he just answered he he just
32:48 answered this question this last week or
32:51 I saw it last week I think it was last
32:53 week because I hadn't seen it
32:55 before where he was he was asked about
32:58 tell me about chat gbt
33:00 5 and you could tell Alman was a little
33:03 exasperated and he was like he was like
33:06 well he
33:08 goes so it's going to get better at all
33:10 the [ __ ] right it's going to get better
33:12 at video and it's going to get better at
33:16 large language model [ __ ] it's going to
33:17 get better at coding it's it's going to
33:19 get better at all the [ __ ] but he goes
33:20 he
33:21 goes mostly what it's going to do is
33:24 it's just going to get
33:27 smarter and like that was his answer the
33:30 answer of what is chat gp5 is it's just
33:33 what if something like chat
33:36 gp4 just got
33:38 smarter well then it would start doing
33:42 things like not hallucinating as much
33:45 being able to do math problems being
33:47 able to code at a high level right like
33:50 like just it getting smarter means it's
33:54 not quite so
33:55 janky and and at what point
33:59 like so you got your janky meter right
34:01 so right now let's say say this is
34:04 completely incapable and this is better
34:07 than every human right right now we're
34:10 we're kind of here like it's mostly it's
34:12 mostly Jank maybe maybe we're over here
34:15 right like it's pretty good it's pretty
34:17 good but there's like a lot of Jank
34:19 there's a lot of improvement to
34:22 go and the thing that's weird right now
34:26 especially for people that are sitting
34:27 on the sidelines of AI is they're going
34:30 oh well these tools aren't great look at
34:32 this Gap this gap's
34:34 huge well if they just get
34:39 smarter it just it's that needle just
34:41 sort of pins and there's a point at
34:43 which I feel like there's a Tipping
34:46 Point where it's good enough that it's
34:49 better than most people and it's
34:51 probably not all the way at it's [ __ ]
34:53 brilliant right it's probably somewhere
34:55 between where we are and where it's
34:57 going to be somewhere in
35:00 there is holy [ __ ] it can do highle
35:05 coding now holy [ __ ] it can write an
35:07 article fact check it grammar check it
35:10 match it to your tone and present it to
35:13 you for just approve and publish or
35:16 would you like to go in and fix anything
35:19 and then you read it and you're like
35:20 [ __ ] that's good
35:24 publish so we'll see we will see but
35:27 we're not there yet
35:29 ethical issues yes absolutely I can't
35:31 access Claude because I live in Greece
35:32 oh that sucks claude's pretty good uh if
35:35 you haven't tried Angelique if you
35:37 haven't tried perplexity you should try
35:39 that too it's pretty
35:44 good hello happy people good day to you
35:47 good day good day good
35:52 day Pure
35:54 Heart wow what a big prediction it will
35:57 get better give that man all the money
36:00 but Pate he only wants $7
36:02 trillion it's not like he's asking for1
36:07 trillion I think here's the thing pay I
36:10 think you're absolutely right I think
36:11 you're on to something
36:13 here if you want to ask for a million
36:16 dollars you got to have all the
36:18 answers if you want to ask for a billion
36:21 dollars you got to have a vision if you
36:24 want to ask for7 trillion doll you just
36:27 say oh the next version's going to be
36:29 slightly
36:34 better I think it's it's called the
36:37 inverse proportion law to the hubris of
36:40 how of how much you're asking
36:45 for listen these these computers are
36:48 pretty smart but if you give me 100
36:52 billi trillion dollar five 5 to 7
36:56 trillion if you give me7 trillion
36:59 they'll get so much
37:01 [Laughter]
37:03 smarter the current code the chat GPT
37:06 recommended to me was awful I'll wait
37:09 for this
37:10 Evolution the current code that it
37:13 recommended to you was awful what do you
37:15 mean recommended to you if you're using
37:18 chat GPT like you use Google like give
37:20 me some code and it gives you code and
37:22 then you judge cat GPT based on that
37:26 interaction you're
37:29 you can get much better results than if
37:31 you're doing it that
37:33 way because it's not just recommending
37:36 code to you like to use it right you
37:38 have some interaction with it you give
37:40 it enough
37:43 context have it analyze its own code be
37:46 in a conversation with it and are you
37:48 using chat GPT 3.5 or chat GPT 4 any
37:52 experience with chat with GitHub
37:54 co-pilot not personally um but a lot of
37:58 of developers I know use it and they I
38:00 like I've heard anywhere from like the
38:03 good developers I know say it's like a
38:05 20 to 30% Improvement in their
38:07 productivity the shitty programmers that
38:10 I know the lazy ones they say it's like
38:13 a 60 to 70% in their productivity so if
38:16 you have a kickass coder they're still
38:19 going to get improvements but if you
38:21 have a mediocre coder they're going to
38:23 get better improvements but yeah but
38:25 it's apparently quite good mod I got
38:30 this just it's 7 trillion yes Source
38:34 Camp it's seven it's 5 to 7 trillion so
38:37 to be clear Sam's not asking for7
38:39 trillion do to to make chat GPT like
38:43 it's not for the software company or
38:44 even the the the compute to buy from
38:48 Microsoft it's to build a chip
38:51 company that dwarfs
38:55 Nvidia so Nvidia right now their market
38:59 cap is is like more than a number of
39:02 countries gdps put
39:05 together because the company was asleep
39:08 at the wheel and they basically seeded
39:10 GPU manufacturing to a single
39:13 company and when this AI [ __ ] took off
39:16 they were the only company making chips
39:18 that you know and had the software that
39:21 that the community was using so what Sam
39:23 alman's saying is if if we want to so so
39:26 here's the deal apparently this is this
39:29 is from what I understand from the geeky
39:32 folks I I I have talked
39:35 to to
39:37 do the quality of video that the the
39:41 Sora
39:43 stuff the you know the Sora video that
39:45 everyone's losing their minds over it's
39:48 apparently 16 A1 100s of GPU power and
39:53 it's and it's like to render a minute of
39:55 video is an hour hour of rendering time
39:59 on 16 A1 100s so I think it's like four
40:03 or five bucks a minute of video which if
40:08 you're used to Runway ml you're like I
40:09 thought that was
40:11 expensive um I don't know if it's going
40:13 to be that I don't like that could be
40:15 [ __ ] but I think it's pretty pricey
40:18 so I think
40:19 Sam's 7 trillion is why don't we make
40:22 some specialized chips that are
40:24 specialized for the way we do things and
40:27 then let's increase production of those
40:30 beyond the world capacity right now it's
40:34 the Cuda ecosystem that really sets them
40:36 apart I don't know if you're serious
40:38 about that or if um if that's a
40:49 joke you need to build context for code
40:52 recommendations yeah
40:54 exactly exactly s hustle Mimi you can't
40:57 just go listen if you're using chat GPT
41:00 like Google oh it's expensive to do
41:02 compute oh oh I see yeah yeah yeah yeah
41:04 yeah exactly it's expensive to do comput
41:06 so so if you look at if you look at
41:10 actually if you join the AI salon and go
41:12 in the mechanics Guild and talk to Pate
41:15 about tpus versus gpus and then if you
41:18 go play with if you go to gro.com
41:21 gq.com right now
41:24 gq.com not the same grock as Elon Musk
41:27 chatty MC Chat
41:29 thing it's it's a tool that looks like
41:32 chat
41:33 GPT it's not a software demo it's a
41:36 hardware demo so if you go to gro.com
41:39 right now and just do a just do a chat
41:42 GPT like thing you'll see how fast it is
41:45 it's like 500 500 plus tokens per second
41:49 so it's like 350 375 words a
41:52 second it's doing the the
41:56 response um
41:59 that's a hardware demo they're demoing
42:02 their crazy efficient chips that are all
42:06 they do is large language models they
42:08 can't do diffusion models they can't do
42:11 they can't make your game faster what
42:13 they can do is they can make large
42:15 language models super
42:17 fast
42:20 so if Sam Alman wants to take that $7
42:22 trillion and design a whole Suite of
42:25 chips all dedicated to the specific
42:27 kinds of compute that that they need as
42:30 efficiently as possible then all of a
42:33 sudden compute costs drop dramatically
42:36 and everybody benefits it also makes
42:39 them you know they they they
42:42 become like that becomes a competitive
42:45 insurmountable there that was
42:49 weird
42:51 huh imagine Sam practicing in the mirror
42:54 saying with a straight Fe face I need $7
42:56 trillion please I I thought it was I
43:00 thought it was [ __ ] Insanity that he
43:02 was saying he needed a 100
43:05 billion that was just the rumors that I
43:07 mean that was just the reporting was
43:09 that you know when he got fired he was
43:10 over asking for a hundred billion dollar
43:14 after Microsoft gave him
43:20 10 yeah it's it's crazy it's crazy
43:23 amount of scale but we'll see we will
43:26 see I'm serious Nvidia didn't make
43:29 objectively better hardware for a while
43:32 but their libraries are better yeah
43:33 that's that's what I've heard PA is that
43:35 the the developers I know and um George
43:40 Hots if you know him was talking about
43:42 that like he George Hots tried to do
43:46 some AI stuff with AMD
43:48 chips and couldn't because the software
43:52 sucked the software libraries were bad
43:56 so yeah it's not that stuff's better
43:58 it's that their their software
44:02 worked uh it's nothing more than a
44:04 probabilistic model it doesn't think yet
44:09 42o
44:10 okay that that dismissive it's nothing
44:13 more than a probabilistic model keep
44:16 keep using it the the more you use these
44:19 tools yes they are probab probabalistic
44:22 model it doesn't mean they're just
44:24 paring back the [ __ ] they've seen it
44:27 doesn't mean They're copying and
44:30 pasting and the minute you add reasoning
44:33 on top of the probabilistic
44:35 model the dismissiveness that is in that
44:39 phrase it's just a probabilistic
44:43 model yeah exactly I understand but how
44:46 do you define thinking
44:51 exactly who was it today someone today
44:53 was just talking
44:55 about some big scientist
45:00 today was just thought was just talking
45:02 about
45:05 this that he said that the way the
45:08 probabilistic model works is essentially
45:11 how our brains
45:13 work he says by definition It's
45:17 thinking the only thing it's not doing
45:19 right now is reasoning the minute it
45:21 does reasoning go look at the David
45:24 Shapiro video on qar go look up qar and
45:26 look at the the diagrams on how that
45:28 [ __ ]
45:30 works it's really cool the the I don't
45:35 know if I have a visual in my head I I
45:36 could probably go find the graphic
45:39 but the way the way qar
45:42 works
45:44 is so you got a large language model
45:47 right and it's it it's got a probability
45:50 and it let's say it can go in one of
45:52 four directions right so you put in a
45:54 prompt and it's got a probability that
45:56 it can go here here here here so what
45:59 how qar works is it says okay I'm my
46:02 goal my goal is over here and I'm here
46:05 and I'm going to take one of these four
46:07 actions Boop and it takes the action and
46:11 then it measures what was the quality is
46:13 the quality of the action that it took
46:15 in the direction of that thing or
46:18 farther away from it and if it's in the
46:20 direction of it it gets a higher score
46:23 and if it's in the other direction it
46:25 gets a lower score so the way I think
46:28 about it is is very much like card
46:30 counting if you've ever learned about
46:32 Blackjack card counting there's a system
46:34 where if it's you know tens are face
46:37 cards it's plus one and if it's uh
46:40 whatever seven or lower it's minus one
46:43 or something like that and all you're
46:44 doing when you're card counting is
46:46 you're keeping track are you positive or
46:48 are you negative and is it you know
46:50 above the Delta right is it like + 5 +
46:53 10 - 5 - 10 and so that's how Q is
46:57 working so it's it's trying to move
46:59 forward toward a goal and it gets you
47:02 know better better better that score is
47:04 going up it's in the right direction
47:05 then all of a sudden it hits a wall and
47:07 now it has to like go around that wall
47:09 and so it's just it's just so that's
47:11 what it's doing none of the current [ __ ]
47:14 none of the current llms are doing that
47:16 at all the current llms are just going
47:17 here's your answer Splat take it or
47:20 leave
47:22 it so the minute it can figure out if it
47:25 if what it did was good or what it's
47:28 doing is in the right direction the game
47:30 changes cuz then it can start doing [ __ ]
47:32 like
47:36 math first you need to make a system to
47:39 understand the idea of a sentence Oh Boy
47:42 mm
47:44 okay $7
47:46 trillion oh that's hilarious is that
47:49 true shop pan I wouldn't I wouldn't have
47:52 met my wife without a that's awesome
47:55 that's good Brandon
47:59 joker7 trillion will buy wait two is it
48:04 billion yeah uh yeah 233 billion 333
48:08 milon 333,000
48:11 333 h100 gpus yeah a trillion dollars is
48:15 a staggering amount of money7 trillion
48:18 is
48:20 like what are they going to do buy a
48:22 [ __ ] country and just take it pave it
48:25 and put the the chip plant on
48:29 it we're going to buy
48:31 [Laughter]
48:35 England we'll use all that uh the tidle
48:38 power we'll buy Amsterdam and use all
48:41 their their title technology to power
48:43 our our chip manufacturing facility I
48:47 don't know what they're going to do it's
48:48 crazy it's crazy crazy
48:56 caaaa
49:00 [Music]
49:14 all right what do you want to do what do
49:16 you want to do here we can we can go
49:17 look at [ __ ] are you in New York City
49:19 did you make an AI assistant app like 10
49:21 years ago um I was in New York City for
49:25 lots of years I was a co-founder of
49:26 agency
49:27 back in the mid90s I
49:30 left I sold agency.com in
49:34 2002 moved out of the city in like 2003
49:41 maybe and then left for Left I moved
49:44 north of the
49:46 city and then moved out to Denver or to
49:49 I moved to Boulder first moved to
49:51 Boulder in
49:54 2011 and I've been out here ever since
49:56 I'm in Denver right
49:58 now um I don't think I made an AI
50:01 assistant I started so I've been doing
50:04 my current company's called storyvine
50:06 which is an automated um video
50:09 storytelling platform we've been around
50:10 for 11 and a half
50:12 years so I don't think
50:15 so I don't think
50:19 so it's not a bad idea Brandon I might
50:22 do
50:26 that
50:29 if Chat gpt's math is right it it
50:34 analyzed well if if it depends on the
50:37 math if Chat gpt's math if chat GPT is
50:40 using Code interpreter and it wrote Good
50:43 python code then it can do
50:45 math but sometimes code interpreter what
50:48 what I found with code interpreter is
50:51 you'll upload some data to it it
50:53 loads um it loads the data from the your
50:58 CSV file into a data frame in in the
51:01 virtual in the in the uh
51:04 container and then it writes python code
51:07 to
51:08 analyze the code that's in the data
51:10 frame but what I've noticed is that
51:14 sometimes it like puts other [ __ ] in the
51:16 data frame that it just makes up or
51:19 it'll just drop a whole chunk of [ __ ]
51:20 out of it and so then it's analyzing
51:23 stuff you're like but wait where's the
51:24 other numbers and it was like oh I'm
51:25 sorry yes I got that wrong let me try
51:27 again and it's just it's just wrong
51:29 again so it's just like I just the the
51:33 reasoning capability of chat GPT right
51:35 now is just stupid so so
51:38 just assume it can't do math oops got
51:41 cut off could 7 trillion do buy set
51:44 up a satellite to manufacture chips you
51:49 mean a satellite like up in the up in
51:52 the planets to Manu I I don't know if
51:54 you gain anything by Manufacturing in
51:56 zero gravity but maybe but $7 trillion
52:00 could buy you a lot of [ __ ]
52:02 satellites
52:05 yes but I assume what we're talking
52:08 about here is big giant sprawling you
52:10 know Tesla scale
52:19 factories 7 trillion won't fit in a
52:22 calculator can chat GPT analyze a
52:24 website and give Su suggestions on how
52:26 to improve it it can can do that all day
52:28 long it's actually good at
52:31 that
52:34 um and Gemini Pro 1.5 I don't have
52:39 access to it yet but the people that
52:41 have got access to it right
52:43 now they're doing things like recording
52:47 screen sessions of them
52:49 using applications and having Gemini
52:53 watch the video and suggest um process
52:56 improvements so I think that I think
52:59 that once you once you start to get like
53:02 million token context windows with
53:04 multimodal stuff you're going to be able
53:07 to use all sorts of inputs to do all
53:09 sorts of crazy [ __ ] but yeah chat jpt
53:12 right now is good you can give it code
53:14 and tell it to suggest how to improve
53:16 things you can give it code and tell it
53:18 to change it and you know make it look
53:22 better or different um you can probably
53:25 just upload an interface and have it you
53:28 know describe it and maybe even describe
53:30 ways to improve it improve
53:33 usability yeah it's good at that stuff
53:36 good good good is it okay to have the
53:39 humble opinion the difference
53:44 between wait is it okay to have your hum
53:48 oh humble opinion the difference
53:51 between an
53:55 employee and an employer with AI I don't
53:58 quite understand the
54:03 question if if you're
54:08 saying can AI completely replace humans
54:12 right now in most jobs the answer is
54:16 no
54:19 but a lot of businesses that I know that
54:23 are starting to get their heads around
54:26 AI
54:28 are looking at their employees saying
54:30 who's AI
54:32 literate and if they're making if
54:35 they're prioritizing employees they're
54:37 prioritizing toward the AI literate
54:42 employees now depending on the job as
54:44 these tools get
54:47 better the ability to completely replace
54:51 a role or function of an employee today
54:55 is going to be very possible
54:58 I think I think it's something like I
54:59 heard this stat today that like 80% of
55:02 the
55:03 jobs like the data jobs that people do
55:06 today are
55:09 repetitive
55:11 um like low contribution just like
55:14 repetitive Cog in the wheel kind of
55:18 tasks that's the perfect kind of thing
55:20 that AI is going to be able to automate
55:24 first so the people that are getting AI
55:27 literate who are expanding their skill
55:30 set they're they're going to be in much
55:32 better shape than people who are sitting
55:34 on the sidelines that this channel is
55:36 all about we can debate whether AI is
55:40 good or bad evil or
55:43 genius um janky and unusable or
55:48 brilliant and [ __ ] profound we can
55:50 debate that [ __ ] all day every
55:53 day and the AI is still coming it's not
55:57 going away all of these peripheral
55:59 conversations about the validity of AI
56:02 and what's it going to do and is it good
56:05 or bad that it replaces shitty
56:07 repetitive
56:09 jobs is all those are all independent
56:13 conversations from the fact that it's
56:17 coming and quite frankly it's
56:21 here it's just that not that many people
56:23 know it yet and the tools are not that
56:27 they're not at the
56:29 level that it's a threat on the
56:35 surface would you ever let an AI robot
56:38 operate on you [ __ ] yes are you kidding
56:43 me did go cap Captain mut go read about
56:49 the robot that they built that Elon Musk
56:53 had to build for
56:55 neuralink it's insane what it does so
56:58 they drill a hole in your
57:02 skull and then in your brain you got
57:05 your
57:06 neurons and you got your little blood
57:09 vessels
57:12 right and what neuralink is is 60 or 120
57:16 I for I don't I forget how big the
57:17 bundles are at this point I think
57:18 they're
57:21 60 there's
57:24 60 um
57:27 what are they called uh little wires
57:31 little probes there's 60 little probes
57:33 that have to go in your brain to connect
57:35 to different neurop Pathways
57:38 right they have to insert them in
57:41 between the blood vessels so that they
57:44 don't poke into one and give you a
57:46 stroke
57:48 so so and and the hole is like I don't
57:52 know the size of a I don't know they I
57:55 think they were originally going to be
57:57 the size of a pencil eraser I think
57:58 they're now about the size of a quarter
57:59 but whatever it's a small hole and so
58:02 this robot looks into your brain and it
58:05 actually steers those things
58:08 around your blood vessels like 60 of
58:11 them and just inserts them one by one
58:13 steering them around the thing so yeah
58:15 would would I let something with that
58:18 level of precision operate on me rather
58:20 than some guy that maybe had too much
58:22 coffee or not enough or went out on a
58:24 bender the night before
58:27 sure yes do I want to be the first guy
58:30 operated on by that robot
58:33 no but by the time a a non-clinical
58:38 trial person is going to be operated on
58:40 by that robot it'll have gone through a
58:42 bunch of [ __ ] I had surgery in early
58:44 2023 with a da Vinci in room physician
58:48 and a remote co-pilot in Florida yeah
58:51 exactly absolutely I
58:55 would absolutely like yes
59:00 yes I mean this is no different to me
59:03 than what what I would I like a
59:07 uh you
59:09 know would I find it acceptable to talk
59:12 to a robot if the robot actually
59:14 answered the phone and could solve my
59:16 customer service problem yes yes I would
59:22 yes I would I would gladly talk to a
59:25 robot that's like hi Kyle how can I help
59:27 you today oh well I tried to open my
59:30 refrigerator door and it fell off in my
59:32 hand what the [ __ ] am I supposed to do
59:34 now oh that's really unfortunate Mr
59:36 Shannon that sounds that sounds like
59:39 something we got to get a technician out
59:41 there right away don't we well yeah if
59:43 you don't want my food to go bad all
59:45 right let's go ahead and do that that
59:47 technician's been booked they're on
59:48 their
59:49 way would I prefer that experience to
59:52 the one that I went through when that
59:54 actually happened yeah
59:57 cuz that was three weeks of [ __ ]
1:00:01 hell did the surgeon press shift enter
1:00:05 exactly that was a joke from last night
1:00:07 there's a there's a kind of funny uh ux
1:00:12 experience on a Microsoft co-pilot and
1:00:16 the neighboring Little Thing Called
1:00:17 Microsoft
1:00:19 notebook which if you didn't see it last
1:00:21 night I'll let you go Discover it for
1:00:22 yourself it's
1:00:25 special
1:00:27 but yeah [ __ ]
1:00:29 yes but I also have a uh I have a world
1:00:34 ID like what's a world ID Kyle I don't
1:00:37 know what a world ID wait
1:00:39 what's
1:00:41 what the hell was
1:00:46 that that was
1:00:49 bizarre it said your live will end in a
1:00:51 minute but I had to like put a puzzle
1:00:54 piece weird
1:00:57 I spoke with gp4 for the first time
1:00:59 today it's often awesome oh that's great
1:01:02 Apple user oh yeah you've been in here
1:01:03 for a while yeah I'm telling you gp4 is
1:01:06 just a different thing I got on co-
1:01:09 co-pilot today and did the exact same
1:01:11 thing so frustrating yeah so if you
1:01:13 don't know in co-pilot right now there's
1:01:16 two modes there's co-pilot you go to co-
1:01:17 pilot.
1:01:19 microsoft.com and then there's co-pilot
1:01:22 and that's just like chat
1:01:24 GPT and you type in a prompt and you hit
1:01:28 return and then if you want to put a
1:01:30 line break in your prompt like if you
1:01:32 want to go hey based on this um thing
1:01:36 I'm about to paste in here shift return
1:01:39 two line breaks and then you paste in a
1:01:40 blob and then shift return and you give
1:01:43 it the rest of your prompt and then you
1:01:45 hit return and it enters it right next
1:01:47 to co-pilot is this thing called
1:01:50 notebook and notebook is a hor it's like
1:01:53 a side by side like instead of your
1:01:54 prompt being on the bottom your prompt
1:01:56 is on the the side and that
1:01:59 one you when you hit return it gives you
1:02:03 a line break and when you hit shift
1:02:05 return it enters it it's the complete
1:02:07 opposite action of the thing that's
1:02:09 right beside it it's Inc it's it's
1:02:12 Insanity tiktock did that to me today in
1:02:15 my life oh good uh investor friendly
1:02:18 agent because that was crazy I thought I
1:02:20 might have been getting booted out of
1:02:21 here forever cuz I
1:02:24 said whatever something something
1:02:29 stupid you can make money
1:02:34 [Music]
1:02:44 with I want customer service like the
1:02:46 Alaskan Airways where their chatbot made
1:02:49 a new policy for bereavement I didn't
1:02:51 hear about that one new security feature
1:02:54 probably because of AI
1:02:57 and people making deep fakes wow yeah
1:03:00 maybe that's
1:03:02 possible but when we get agents because
1:03:05 these things have Vision all of these
1:03:07 captas and Slide the puzzle piece to the
1:03:10 puzzle piece all of the
1:03:13 AIS so the AIS are going to be able to
1:03:16 do all that so there's going to have to
1:03:18 be a completely new method like all of
1:03:21 the current security two Factor
1:03:24 authentic all that bullshit's going to
1:03:25 just
1:03:26 get obliterated in the next three or
1:03:28 four years
1:03:30 so it's going to be crazy it's going to
1:03:32 be I'm I'm telling you 2023 last normal
1:03:38 Year all right oh that was Canada air
1:03:41 and it was actually eight eight months
1:03:43 ago got my big presentation about school
1:03:45 AI tomorrow Mr K good luck with that so
1:03:48 everybody uh tap the screen wish wish
1:03:51 wish Mr K good luck so Mr K is a teacher
1:03:55 in central
1:03:56 Pennsylvania I grew up in York
1:03:58 Pennsylvania so East Eastern York High
1:04:02 School Mr K so just south of you there
1:04:05 buddy that's that's how you got this
1:04:07 [ __ ] genius pile of gray
1:04:11 matter
1:04:14 um I guess pitch into the school board
1:04:16 tomorrow about um we should be using AI
1:04:20 what Mr K is going to go tell them is um
1:04:23 Hey
1:04:24 listen there's these things out there
1:04:27 called large language
1:04:29 models and there are students our
1:04:33 students are large language models and
1:04:35 it's our job to get language in their
1:04:39 frustrating little
1:04:41 brains and I don't know if you've heard
1:04:43 this but another thing came along called
1:04:45 a large language model that are these
1:04:47 computer programs that are the most
1:04:49 remarkable knowledge systems in the
1:04:52 history of humanity and right
1:04:55 now we're not allowed to use
1:04:58 them and so what I would like to suggest
1:05:01 to us as a community is maybe we should
1:05:05 take these large language models and
1:05:07 help these large language models learn
1:05:09 faster and better and different as if
1:05:13 these new things were going to be around
1:05:16 when these kids
1:05:18 graduate so wish Mr K good [ __ ] luck
1:05:23 yes go tell them go blow their minds and
1:05:27 go tell them this [ __ ] isn't going
1:05:31 away these are not plagiarism machines
1:05:34 these these are not cheating
1:05:37 machines these are Humanity amplifying
1:05:42 tools our kids will learn faster deeper
1:05:49 better they will be more passionate
1:05:51 about [ __ ] their education will be more
1:05:55 personalized
1:05:57 holy [ __ ] [ __ ] you don't want me in
1:05:59 the
1:06:02 room like listen
1:06:11 dipsticks so
1:06:13 please if you're in the AI Salon go into
1:06:16 the education Guild
1:06:18 and give Mr K some shout outs and some
1:06:21 good loving education Guild Monday night
1:06:23 meeting right before the AI learning
1:06:26 awesome fantastic thank you Brandon
1:06:28 that's great so I keep talking about the
1:06:31 salon what the hell is that
1:06:33 so if
1:06:37 you are curious about
1:06:40 AI or if you're kind of like me where
1:06:43 you're like I kind of think this AI
1:06:44 thing's a big deal but everyone else in
1:06:46 my life seems to think it's not a big
1:06:48 deal am I wrong no you're not wrong
1:06:51 you're just hanging out with the wrong
1:06:54 crowd you're hanging out with bad
1:06:57 influencers love that Humanity
1:06:59 amplifying Kyle for president
1:07:02 2028 this
1:07:05 channel so okay so if you don't know if
1:07:07 if you're new
1:07:08 here there's a pile of people in here if
1:07:11 it seems like there's people in here
1:07:13 that kind of know each other and have
1:07:15 inside jokes and are answering your
1:07:17 questions those are the people that show
1:07:19 up here
1:07:21 regularly so if you don't know I do
1:07:23 these things seven nights a week
1:07:26 why because I know how to make Good Life
1:07:31 Choices the Irregulars the ones that
1:07:34 show up here every
1:07:37 night are living proof that this AI
1:07:42 stuff are Humanity amplifiers because to
1:07:45 a person the people that just show up
1:07:47 here regularly and are in this
1:07:49 conversation in what conversation it
1:07:52 doesn't seem like you've said anything
1:07:53 tonight yeah exactly you know why
1:07:55 they're here cuz while I'm rambling
1:07:57 they're playing with
1:07:59 AI and if I mention some tool they go
1:08:01 Google it oh that's kind of cool then
1:08:02 they're playing with it and they're
1:08:03 playing with they're doing this and
1:08:04 they're doing that and they're like oh
1:08:06 [ __ ] all of a sudden oh I can make a
1:08:08 really good image and then they share
1:08:09 that image in the AI salon and then
1:08:12 someone else sees that image they're
1:08:13 like how did you do that like oh I don't
1:08:14 know I I uploaded this image and then I
1:08:17 had it do that and then oh my God that's
1:08:19 amazing and then they're talking to one
1:08:20 another and and and what's happening
1:08:24 is they're getting
1:08:27 reignited creatively like like you know
1:08:31 the human spark that spark that's like I
1:08:34 got something to say I I want to put
1:08:36 something into the world you know that
1:08:38 thing you had when you were
1:08:39 seven and then the world [ __ ] beats
1:08:42 it out of
1:08:45 you people that are playing with these
1:08:47 tools regularly that like that that fire
1:08:50 is reigniting and they're like oh what
1:08:54 what if I did this
1:08:56 what if if I did that could I do this
1:08:59 and so I I get to witness it in here
1:09:01 every night everyone who shows up here
1:09:03 every night I'm The Quiet One a newer
1:09:05 regular awesome yeah jimc I know you've
1:09:07 been in here a lot it's
1:09:10 great is we're all getting to witness
1:09:13 and support and celebrate one another as
1:09:16 we're rediscovering who we
1:09:19 are that's what these tools
1:09:23 are can they be used for bad might they
1:09:26 do bad things
1:09:29 yeah in the
1:09:32 meantime until it shuts down the grid
1:09:35 and ruins our lives
1:09:37 forever why don't we see what they are
1:09:40 why don't we play with
1:09:42 them that's what this group
1:09:47 is so if you go there it's going to take
1:09:49 you to a link tree you can there's merch
1:09:51 there you can buy T-shirts that support
1:09:53 this group because this doesn't cost
1:09:55 anything
1:09:57 there's a similar Channel out there that
1:10:00 kind of does what I do he created a
1:10:02 community he's charging 500 bucks for
1:10:06 it this is
1:10:09 free not like free to get into level
1:10:18 one and the people that are
1:10:22 here are [ __ ]
1:10:24 remarkable thank thank you Sherry D
1:10:27 thank you so much Sherry D is one of
1:10:29 them 30-year nursing career she's now
1:10:33 making music
1:10:34 videos for
1:10:36 bands like really good ones that's
1:10:39 exactly what's happening I felt like job
1:10:42 the job market is trash and this gives
1:10:43 me hope
1:10:45 yep exactly
1:10:47 riggity
1:10:50 exactly and what's the hope I don't know
1:10:54 what's this thing going to give you
1:10:56 that's going to be the thing that's
1:10:57 going
1:10:58 to help you survive all this upheaval I
1:11:01 don't know but here's what I know if
1:11:03 you're sitting on the sidelines with
1:11:04 your arms
1:11:05 crossed these tools ain't all that
1:11:09 you're not going to be in as good a
1:11:10 shape as someone who's hanging out in
1:11:12 here figuring out what the hell this
1:11:13 stuff can
1:11:14 do promise you
1:11:16 that so if you go there the first link
1:11:19 says the AI Salon Community click on
1:11:21 that and it's going to take
1:11:24 you
1:11:31 to
1:11:35 here welcome to the AI
1:11:39 salon and you know all your friends that
1:11:41 are like you know Sally
1:11:47 um we need to talk so here's the thing I
1:11:50 know we said we wanted to go out to
1:11:51 drinks tonight but this is kind of a
1:11:53 little bit of an intervention I don't
1:11:55 have a drinking problem no it's not
1:11:56 about the booze we like no we like
1:11:59 drinking it's about the AI could you we
1:12:03 we we've talked and we've decided that
1:12:05 you talk about AI it's all you talk
1:12:08 about anymore and um how do we say this
1:12:13 um we want the old Sally back um could
1:12:16 could we just do some melon poppers
1:12:19 [Laughter]
1:12:22 and so if your friends are doing that to
1:12:25 you
1:12:27 this group is not
1:12:29 that this group is like oh you're really
1:12:32 into AI cool come
1:12:35 on oh you're really curious about it
1:12:37 cool come
1:12:38 on and notice this is not a single
1:12:41 person heading off into the abyss one of
1:12:44 the one of the realities right now I
1:12:46 promise you I promise you to a person
1:12:49 every single person in the AI salon
1:12:51 right now feels like they are behind and
1:12:54 undereducated and UND skilled like like
1:12:58 every someone else in this community has
1:13:01 to know more than they do every single
1:13:03 person feels like they can't keep up
1:13:06 they don't they're not good at this
1:13:09 [ __ ] they like the stuff that they're
1:13:11 doing but they apologize for it because
1:13:13 it's not
1:13:14 enough to a person I'm trying to
1:13:17 eradicate the apologizing in this group
1:13:19 by the
1:13:20 way because what the people are doing in
1:13:22 this group is remarkable because what
1:13:24 they're doing is they're
1:13:28 playing and that's the only skill you
1:13:30 need right
1:13:32 now so this group will welcome you in
1:13:36 with open arms and then here's my
1:13:38 request so if you're
1:13:39 here read about who we are and why we do
1:13:42 what we do and then what I really want
1:13:43 you to take in like I really want you to
1:13:45 take this in like like as passionate as
1:13:48 I'm talking right now I want you to
1:13:51 really read the
1:13:53 values so the values here
1:13:56 are not like this is who we want to be
1:13:59 the values here of this is who this
1:14:01 community already
1:14:04 is and the thing that's happening here
1:14:06 is this community is starting to to
1:14:09 grow quickly and and it's it's starting
1:14:12 to
1:14:13 accelerate and I've grown you know I've
1:14:17 start I started a company that grew
1:14:19 really big really fast I started a an
1:14:21 organization kind of like this one in
1:14:23 1994 called the worldwide web artist
1:14:25 Consortium That Grew really big really
1:14:28 fast and one of the things that happens
1:14:30 when you grow really big really fast is
1:14:32 you you kind of
1:14:35 lose the good [ __ ] cuz the new people
1:14:39 overwhelm the old people and just it
1:14:41 just becomes
1:14:42 chaos so the reason I talk about these
1:14:45 values most nights I I think every
1:14:49 night is that you shouldn't join this
1:14:52 community if you don't dig these values
1:14:55 because this this is who this community
1:14:56 already
1:15:02 is okay so the first one's curiosity
1:15:04 like we want people in here that are
1:15:06 curious not cynical right you can be
1:15:08 cynical but just be funny about it like
1:15:10 if you're cynical like a standup
1:15:12 comedian is
1:15:13 cynical like like Tom
1:15:18 sigura or or Doug stanh hope cool that's
1:15:21 fine but in here you should be curious
1:15:24 about this AI stuff huh H wonder how
1:15:26 that works whoa what would happen if I
1:15:28 did this right next
1:15:31 next next value is is
1:15:34 generosity so don't be stingy with like
1:15:37 oh I just learned something I'm going to
1:15:38 keep that to myself no no I just learned
1:15:40 something let me see if anyone else
1:15:42 knows that too here's how I did this
1:15:44 anyone else oh how'd you do that oh
1:15:46 here's how I did that a lot of back and
1:15:48 forth right
1:15:50 generosity
1:15:52 exploration people that are excited to
1:15:54 like strap on the backpack and just head
1:15:56 into the unknown cuz that's all what
1:15:59 we're what we're all going to be facing
1:16:01 for the next 5 years is a lot of
1:16:05 unknown so you want to be around
1:16:10 people who even if they're not
1:16:12 comfortable in that scenario are willing
1:16:15 to do it this is a community of people
1:16:18 willing to step into the
1:16:22 unknown do you know how [ __ ]
1:16:24 inspiring that is to be hanging out with
1:16:26 people that are willing to do
1:16:27 that you know how few people in life are
1:16:30 willing to do
1:16:33 that so so that's who this group is
1:16:35 collaboration is another one you know
1:16:37 willing to do it together
1:16:39 inclusion like I don't care if you've
1:16:41 got you know six hours experience with
1:16:44 AI or six decades experience with AI
1:16:47 doesn't matter all points of view right
1:16:50 now are contributing to everyone's
1:16:53 understanding because no one knows
1:16:54 anything everything's changing too
1:16:57 fast empathy is a big
1:16:59 one people are scared people are angry
1:17:03 people are dismissive there's been some
1:17:05 of that in here tonight oh you know
1:17:07 these things are just predictive
1:17:10 engines well they're not perfect well
1:17:13 they can't do high level coding okay
1:17:17 understand but understand that sitting
1:17:20 underneath that kind of
1:17:24 bravado is
1:17:26 fear underneath that
1:17:28 anger
1:17:30 is insecurity oh no what's going to
1:17:33 happen if if if this thing I've been
1:17:35 doing for 20 years all of a sudden is
1:17:37 irrelevant it's very real right so
1:17:40 imagine being in a community of
1:17:43 people who tap into that and are willing
1:17:46 to hear that and not just interact with
1:17:49 your dismissiveness or your anger but
1:17:52 are willing to say oh yeah what's really
1:17:54 going on on here well I've been doing
1:17:57 all this thing and I don't feel like
1:17:58 learning a new career yeah that must
1:18:00 suck come on in let's let's see if we
1:18:02 can figure something out together like
1:18:04 that's who this group is and then the
1:18:05 last one is
1:18:07 gratitude
1:18:09 like actively
1:18:13 actively acknowledging one
1:18:16 another if someone if someone around you
1:18:18 is generous enough to tell you how
1:18:21 they're doing something or give you a
1:18:22 leg
1:18:23 up thank them
1:18:30 the more bulgy I get that's funny Corey
1:18:33 Sandler I'm on the put me on put me in
1:18:36 coach I'm ready to
1:18:41 play oh man so anyway that's the values
1:18:44 of this group and I I know that was a
1:18:46 little a little preachy but [ __ ] it I
1:18:48 mean I mean I'm very serious
1:18:54 about how remarkable this group is and
1:18:57 how much how committed I am to keeping
1:19:00 it that
1:19:02 way um I mostly only use an iPad and my
1:19:05 iPhone can I get into AI stuff with just
1:19:07 that absolutely pres dog so pres doog so
1:19:10 you can right now just using
1:19:14 Safari um go to chat.
1:19:18 open.com I would strongly recommend
1:19:21 subscribing to chat GPT plus which is
1:19:24 their 20 Buck a month
1:19:26 um subscription to upgrade you to chat
1:19:29 GPT 4 you can also download the chat GPT
1:19:32 app make sure you download the official
1:19:35 one from open AI make sure that the
1:19:38 company that makes chat GPT is open AI
1:19:42 anything other than that is a faker app
1:19:44 that's out to rip you
1:19:46 off but yeah you can absolutely do all
1:19:49 this stuff right there I would also
1:19:51 download the PI
1:19:53 app I think perplexity has an app I
1:19:56 would download the perplexity
1:19:59 app I think Claude has an app from
1:20:08 anthropic so yes
1:20:13 absolutely okay so when you come in here
1:20:15 so if you if you dig what what we're
1:20:18 about
1:20:19 cool the next thing I want you to do is
1:20:22 introduce yourself take that Val of
1:20:27 generosity and come in and introduce
1:20:31 yourself mess Schmidt said hi that's a
1:20:34 little
1:20:36 stingy but he at least introduced
1:20:38 himself right I think of
1:20:41 he but ideally what you do is you come
1:20:44 in here you say hey here's who I am
1:20:47 here's what I'm doing here's what I dig
1:20:49 in my life here's what I've done in my
1:20:51 life here's where I am with
1:20:53 AI sure would would be swell to find
1:20:56 someone who blank does what does
1:20:58 something knows something about this is
1:21:01 willing to just help someone out
1:21:06 right and then you can check out our
1:21:11 events and if you go to this this third
1:21:14 link the the two links below introduce
1:21:17 yourself is this thing called AI
1:21:18 Launchpad and
1:21:20 tools and that's a page that I
1:21:22 personally
1:21:24 curate wait progg I'm old 59 just a
1:21:27 truck driver and a pilot but AI
1:21:29 seriously intrigues me pres dog I'm
1:21:33 58 and and I I'm just a I'm I'm just a
1:21:38 guy that that's playing around with uh
1:21:40 with storytelling tools and Technology
1:21:43 the the what you what you do and have
1:21:48 done is irrelevant to what's going on
1:21:50 right now if AI seriously intrigues me I
1:21:53 mean seriously okay okay you're a truck
1:21:56 driver okay PR dog you're going to
1:21:58 [ __ ] love this go hang
1:22:02 on hang
1:22:04 on
1:22:06 go so you only have an iPad and an
1:22:09 iPhone right so you got Bluetooth in
1:22:10 your
1:22:13 rig so I want you to you can do this
1:22:15 with either the chat GPT app or you can
1:22:18 get this app called Pi Pi it's it's Pi
1:22:21 personal assistant and the the icon
1:22:23 looks like that it's it's green two
1:22:27 green letters
1:22:29 Pi go get that
1:22:33 app so it's it's like chat
1:22:38 GPT except it's optimized for
1:22:43 conversation and so oh you have the pie
1:22:46 okay cool perfect so I don't know if you
1:22:48 use it if and I don't know if you use it
1:22:50 like this but go on your long runs to
1:22:55 2hour drive 4our drive 8 hour drive
1:22:57 whatever it is right and just start
1:22:59 having a conversation with pi it's
1:23:01 connected to the internet and it's
1:23:02 designed to talk and you know so there
1:23:06 you are and if you don't know it that
1:23:08 little icon right there the little phone
1:23:09 icon puts it into conversation
1:23:15 mode I'm on a long trip and I'm a little
1:23:18 bored can you help me come up with a way
1:23:20 to pass the
1:23:23 time
1:23:25 sure thing Kyle long trips can
1:23:27 definitely be a bit tedious sometimes
1:23:30 let's see if we can spice things up a
1:23:31 bit how about we play a game of would
1:23:34 you rather I'll give you two options for
1:23:36 different scenarios and you have to
1:23:37 choose one ready to get started now I
1:23:41 think what I'd rather do is you tell me
1:23:43 a a news story that happened today but
1:23:46 keep the details from me and see if I
1:23:48 can
1:23:49 guess about
1:23:53 it
1:23:57 all right Kyle here's a news story that
1:23:59 happened today I'll keep the details
1:24:01 hidden so you'll have to use your
1:24:02 detective skills to figure it out a
1:24:05 major event occurred in a region that
1:24:08 has been in the headlines recently
1:24:10 leading to a significant loss of life
1:24:12 and injuries it involved a large group
1:24:15 of people and a convoy of Aid yeah yeah
1:24:18 I know what that one is I'm not going to
1:24:20 talk about that one I'm not getting
1:24:21 booted off Tik Tok for that but
1:24:23 anyway you can you know it can make up
1:24:27 games for you it can talk to you you can
1:24:29 have deep you know contemplative
1:24:32 conversations with it you can have it
1:24:34 teach you something you can have it
1:24:36 explain books to you movies to you I
1:24:38 just all sorts of stuff so
1:24:41 um that would be a really like having
1:24:46 having pie teach you about AI while
1:24:48 you're driving would be kind of kind of
1:24:52 wild and cool cuz the other thing about
1:24:54 those conversations is you're having
1:24:56 these conversations it's recording them
1:24:58 all in
1:24:59 text the other thing you can do with
1:25:02 chat
1:25:04 GPT is if you've got chat GPT 4 it's got
1:25:09 dolly in
1:25:10 it so you do a similar sort of thing
1:25:14 with chat pt4 but you can have it make
1:25:20 images I'm on a long road trip right now
1:25:23 and I'm driving through Mid Ohio can you
1:25:26 make me a picture of a blue semi truck
1:25:30 driving down the highway in Mid
1:25:39 Ohio and so it's off now rendering an
1:25:42 image so so not only can you have a
1:25:44 conversation you could be making art
1:25:47 while you're
1:25:50 driving now why you would want to do
1:25:52 that I don't know
1:25:55 but you would at least start to learn
1:25:57 the capabilities of these things and
1:25:59 what's
1:26:01 possible and's see what it does see what
1:26:03 it comes up
1:26:04 with look there's here's a picture of a
1:26:07 blue semi- Tru driving down the highway
1:26:09 in Mid Ohio it really captures the sunny
1:26:12 day and the essence of a road trip
1:26:14 through the Midwest how's your road trip
1:26:16 going so far you can check out the
1:26:19 images in our conversation
1:26:22 history it's going okay I'm a little
1:26:24 board though can you come up with a game
1:26:26 for us to
1:26:30 play sure how about a classic game of 20
1:26:33 questions it's perfect for road trips
1:26:36 I'll think of something there you go so
1:26:40 cool
1:26:43 right anyway so
1:26:46 yes yes get curious about it here what
1:26:49 what's what's happening in here pres
1:26:52 dog pretty consistently
1:27:01 is people kind of come in and they're
1:27:03 like I think I want to learn about Ai
1:27:05 and they're like I don't know anything
1:27:08 help and
1:27:11 then very often what they realize is
1:27:14 everyone they talk to feels the same way
1:27:16 even if they've been doing this AI stuff
1:27:18 for a while like people might know more
1:27:19 than them but they're like oh everyone's
1:27:21 kind of trying to figure this out and
1:27:24 there's something really comforting in
1:27:28 that and then if you haven't had it yet
1:27:33 you
1:27:35 know at some point you'll have what I
1:27:38 call your Kevin McAllister moment right
1:27:40 from Home
1:27:42 Alone where these tools will do
1:27:45 something that is so good or profound or
1:27:49 just
1:27:50 unexpected that you're like I didn't
1:27:52 know it could do that right and it's
1:27:56 and then the the more you use these
1:27:58 tools the more you have the sort of
1:27:59 rolling series of Kevin mallister
1:28:04 moments and if you keep pushing and keep
1:28:08 playing and keep sort of figuring out
1:28:10 where the edges of your interests
1:28:13 are all of a
1:28:15 [Music]
1:28:17 sudden you'll discover something you're
1:28:20 either good
1:28:22 at or you were super passionate about
1:28:24 out when you were younger or something
1:28:26 where all of a sudden it like it will
1:28:27 reignite something in you or it it it'll
1:28:30 ignite something new in you and you'll
1:28:32 be like oh well let me explore this let
1:28:35 me explore that and I'm just I see it in
1:28:37 here every flipping night what are your
1:28:39 thoughts on AI and Automation in
1:28:43 homes
1:28:45 um my my thought on any sort of complex
1:28:48 systems like home automation systems are
1:28:51 super
1:28:53 cool
1:28:55 if you're an Uber
1:28:58 nerd right
1:29:02 because all the home automation systems
1:29:05 they've all got their own proprietary
1:29:07 [ __ ] systems and they're all
1:29:09 incompatible with each other and they
1:29:12 all have really shitty
1:29:14 interfaces and so most home automation
1:29:17 nerds started with this system and then
1:29:20 added some components from that system
1:29:22 and then they've got the their Sonos
1:29:24 system over here they got that system
1:29:26 over there and they got their GE LED
1:29:29 dial color light bulbs over here and
1:29:31 none of them talk to each
1:29:34 other and so only the quantum physicist
1:29:37 in the house that put the [ __ ] together
1:29:39 can deal with
1:29:41 it right and so I think what AI is going
1:29:45 to do is it's going to
1:29:47 simplify those complex systems where
1:29:50 when we have say an autonomous agent
1:29:52 that you can talk to like pi and you can
1:29:55 just say hey go figure this [ __ ] out
1:29:58 make it warmer and make it green in here
1:30:01 and turn on the TV and turn down the
1:30:04 radio and make the lights blink
1:30:08 purple and then just just start doing
1:30:12 it then it'll be
1:30:15 cool so I don't know if anyone's built
1:30:19 that
1:30:19 yet but I like I haven't touched home
1:30:23 automation [ __ ] cuz I don't feel like
1:30:25 spending 17 hours a week debugging why
1:30:29 my light won't turn on in
1:30:33 the you know next to the the reading
1:30:38 Nook I had no interest in it but if I
1:30:40 can just say AI figure that [ __ ] out
1:30:42 I'll just go plug some [ __ ] in I'll I'll
1:30:45 give you access to all the [ __ ] I
1:30:47 plugged in and you figure it out then
1:30:49 I'm in I did ask chat gbd 3.5 to find me
1:30:53 cheap flights and actually worked and
1:30:55 that that's actually remarkable cuz 3.5
1:30:57 isn't connected to the internet so how
1:31:00 did it do that you sure or did you use
1:31:06 co-pilot if you want if you want access
1:31:08 to chat GPT 4 the underlying engine gp4
1:31:13 that's connected to the internet but you
1:31:14 can't afford the 20 bucks a month for
1:31:16 chat GPT go to Coop pilot.
1:31:20 microsoft.com I've watched you many
1:31:22 nights I've learned a lot just listening
1:31:24 to you oh that's awesome PR dog I
1:31:25 appreciate it yeah I thought I've seen
1:31:27 you in here
1:31:28 before I [ __ ] love that you're a
1:31:30 trucker and that you're you're trying to
1:31:32 figure this [ __ ] out I I I I just I feel
1:31:35 like anyone who's willing to go into
1:31:38 that unknown uncomfortable space and
1:31:41 just try to figure some [ __ ] out is
1:31:44 going to be
1:31:46 gifted like all sorts of New
1:31:48 Opportunities so I think I think where
1:31:50 you are right now is like
1:31:53 perfect like I'm intrigued by this but
1:31:56 this is what I do like that whatever the
1:31:58 limiting beliefs are of however you've
1:32:00 defined yourself right that that maybe
1:32:03 this is out of your reach drop that [ __ ]
1:32:07 this is this is this is actually really
1:32:08 important November 30th 2022 when chat
1:32:11 GPT launched here's here's the big deal
1:32:14 about it from from where I
1:32:17 sit prior to chat GPT if you wanted to
1:32:21 do this AI [ __ ] you kind of had to be a
1:32:24 geek
1:32:25 you like it didn't hurt if you went to
1:32:27 MIT having the Stamford mathematics
1:32:30 degree probably a good thing if you're
1:32:32 trying to figure out tensor
1:32:35 flow and and embeddings and mathematical
1:32:39 weightings and probability
1:32:42 matrices chat
1:32:45 GPT does for machine learning and AI
1:32:49 what the worldwide web did for the
1:32:50 internet so prior to the worldwide web
1:32:53 the internet was command line it was
1:32:56 researchers it was scientists it was
1:32:59 academics and then the worldwide web
1:33:02 comes along where you can sort of point
1:33:03 and click on hyperlinks and the only
1:33:06 skill you need to be able to use the
1:33:07 Internet is this click click click click
1:33:10 click chat GPT does the same thing for
1:33:14 all this
1:33:15 stuff all you need to be able to do is
1:33:18 talk to it and I think because of terms
1:33:22 like prompt Eng engineering is really
1:33:25 important you've got to be a prompt
1:33:26 engineer it makes it sound really
1:33:29 [ __ ]
1:33:31 intimidating what prompt engineering is
1:33:33 is use your words good to ask for [ __ ]
1:33:37 you want
1:33:39 good and then the machines will give you
1:33:41 good stuff back like that's what it is
1:33:45 now there's all sorts of ways you can do
1:33:47 prompting one of them is you can do
1:33:49 prompt engineering and very structured
1:33:51 prompting and approach it like a systems
1:33:53 analyst
1:33:55 another way you can use prompts is just
1:33:57 talk to these
1:33:59 things as if you were talking to someone
1:34:01 who could do [ __ ] hey could you do that
1:34:05 [ __ ] sure here's that [ __ ] that's not
1:34:07 very good could you do that [ __ ] better
1:34:09 sure here's better that
1:34:11 [ __ ]
1:34:13 like that's
1:34:15 it and if you're willing to just dive
1:34:18 into it and go [ __ ] it I'll just try I
1:34:22 think the rewards are insane
1:34:25 I've been missing these sessions you
1:34:26 have really changed my my AI life thank
1:34:29 you very much thank you Carl welcome
1:34:31 welcome back yeah just keep coming keep
1:34:33 coming keep hanging out this is
1:34:38 um use your words good exactly Lori you
1:34:42 got to use your words
1:34:45 good and you want to know something
1:34:47 [ __ ]
1:34:49 amazing if you're not so good with the
1:34:51 words say to chat GP
1:34:54 hey I want to do a thing but I'm not so
1:34:57 good with the words can you help me come
1:34:58 up with a
1:35:01 prompt that will get you to do this
1:35:04 thing better and guess what it will in
1:35:06 fact when you ask chat gbt to make you
1:35:08 an
1:35:09 image it doesn't use your prompt you
1:35:12 want to see
1:35:14 [Music]
1:35:19 something okay here we are chat GPT chat
1:35:22 GPT 4
1:35:25 we're just in normal chat
1:35:27 gbt I love the way Kyle talks about
1:35:29 playing around to find out and learn I
1:35:31 think it's the only I mean I appreciate
1:35:33 that thank you Winston I think it's the
1:35:35 only the only thing
1:35:37 that does not seem like insanity to me
1:35:40 right now is playing because it it it
1:35:43 actually seems like insanity to me right
1:35:45 now to try to get your head around all
1:35:47 of what's happening and try to master
1:35:49 any of these
1:35:50 tools it's just it's it's stupid because
1:35:53 they're just they're changing too fast
1:35:55 there's too many of them so so don't
1:35:58 take it so seriously like this is a rare
1:36:00 moment in history where where we get to
1:36:02 just play with this [ __ ] okay so I'm in
1:36:04 regular chat GPT I'm going to say I'll
1:36:06 I'll do my classic make
1:36:08 me
1:36:10 a me a
1:36:13 wide photo of
1:36:17 a
1:36:19 green
1:36:21 70s muscle car in an
1:36:27 abandoned
1:36:30 Factory all right so I did those words
1:36:34 pretty good they were pretty good words
1:36:36 they were not
1:36:39 bad they're they're not great words
1:36:42 right green 70s muscle card an abandoned
1:36:45 Factory a wide
1:36:47 photo so we're going to we're going to
1:36:49 let it make its little picture here
1:36:50 we'll see what it comes up
1:36:52 with I love to listen to you thank you
1:36:55 very much I appreciate
1:36:58 that side hustle Mimi have a good night
1:37:00 everyone I'm falling asleep at the keys
1:37:02 good night good night
1:37:06 Mimi it's it is my job I'm like you
1:37:10 could take Xanax or you can listen to
1:37:12 the AI learning lab I'll knock you right
1:37:22 out oh man
1:37:32 come on chat GPT what's going on with
1:37:34 this image show me the
1:37:40 image where my image where my
1:37:44 image I used it to fix my res resume
1:37:48 perfect use
1:37:49 case um let's stop this and restart it
1:37:55 uh I'll do try
1:38:00 again when will emo have an app um I
1:38:04 would imagine emo is so powerful so by
1:38:07 the way emo so if you don't know what
1:38:08 emo is emo is a tool that Alibaba
1:38:12 actually came out
1:38:14 with
1:38:16 um that will take a still image and a
1:38:19 recording of someone
1:38:21 talking and not only animate the lips
1:38:25 but like fully animate the head eyebrows
1:38:28 all the facial
1:38:29 expressions um and it came out yesterday
1:38:32 as a white paper
1:38:34 um Mr Studio it is like it it was one of
1:38:39 those white papers that so many people
1:38:40 are talking about it that I would be
1:38:43 surprised if by the end of the weekend
1:38:46 there wasn't some um Google collab
1:38:50 notebook that you could just you know
1:38:53 get up up and running and have it
1:38:55 running like I would imagine that
1:38:57 there's probably some version of that
1:38:58 right
1:38:59 now um anyway I was talking before about
1:39:03 promp prompting and prompt engineering
1:39:06 and how I I think that's that term is
1:39:09 doing beginners of big disservice that
1:39:12 get intimidated by it and that you can
1:39:15 just use simple words and and these
1:39:18 tools are actually getting better at
1:39:19 helping us so the way chat GPT works I
1:39:23 typed in I created oh wait I I typed in
1:39:26 make me a wide photo of a green 70s
1:39:28 muscle car in an abandoned Factory if I
1:39:30 click on this image in the upper right
1:39:33 hand corner is the information button
1:39:36 here's the prompt that chat GPT
1:39:39 wrote a wide photo depicting a vibrant
1:39:42 green muscle car positioned in the midst
1:39:44 of an abandoned factory setting the
1:39:47 Factory's interior is marked by a vast
1:39:50 desolate space filled with remnants of
1:39:52 industrial machinery
1:39:54 rusted with Co rusted and coated with
1:39:57 Decades of neglect sunlight streams
1:39:59 through the broken painless Windows
1:40:02 Illuminating the dust particles in the
1:40:03 air and creating a dramatic interplay of
1:40:06 light and Shadow the muscle car with its
1:40:08 glossy well-maintained exterior stands
1:40:11 in stark contrast to the surrounding
1:40:14 Decay offering a striking I I mean this
1:40:18 is [ __ ] impressive
1:40:21 right Kyle is my new Night Show theme
1:40:33 [Music]
1:40:39 song um isn't that
1:40:42 crazy so I typed in a shitty
1:40:45 prompt and it said I got you boss here
1:40:48 you go here's an incredibly well-written
1:40:53 incredibly
1:40:54 descriptive prompt that's going to
1:40:58 generate
1:41:02 that so don't be intimidated by this
1:41:05 [ __ ] just play with
1:41:08 it you'll be surprised oh by the way did
1:41:11 you guys
1:41:14 know
1:41:18 sunno sunno AI that does songy stuff I
1:41:23 didn't know this I found this out last
1:41:24 night after I was leaving or after I
1:41:26 left now available V3 Alpha
1:41:32 access for pro and Premier Members I
1:41:35 don't think I'm a pro
1:41:38 member sign in with Google did I pay for
1:41:42 this maybe I paid for
1:41:48 this what's
1:41:50 new V3 access
1:41:55 extended Max song Length faster song
1:42:08 generation
1:42:09 where custom mode
1:42:15 V2 yeah I don't think I've got access to
1:42:21 it but anyway if you haven't seen
1:42:24 I love to oh thank
1:42:27 you
1:42:29 um let's
1:42:32 see
1:42:38 country song about driving my big rig
1:42:44 through Mid
1:42:52 Ohio
1:42:54 yeah I've only got
1:42:56 V2 but anyway there's a V3 of this that
1:42:58 you can do longer song lengths and
1:43:00 apparently it's better so if if you like
1:43:02 making music but man this is much faster
1:43:06 rolling on the mid
1:43:08 Ohio highway let's these these could be
1:43:16 bad kicking
1:43:19 a I hit the RO make my way through the H
1:43:25 nothing going to make me St that's
1:43:28 [Music]
1:43:40 okay it's not singing where's the
1:43:43 [Music]
1:43:48 words and it didn't do the words for
1:43:50 that one that's funny sometimes it
1:43:51 doesn't do words janky is
1:43:54 dances in Ohio
1:43:57 and yeah anyway new version of sunno out
1:44:01 so if you're into the suo stuff go go
1:44:04 get the V3 of
1:44:08 [Music]
1:44:12 it all right
1:44:15 everybody I think I'm going to wind I'm
1:44:18 going to wind her
1:44:22 down
1:44:25 um tomorrow so a couple of things
1:44:28 tomorrow I've got AI office hours on
1:44:32 LinkedIn so if you find my profile on
1:44:34 LinkedIn if you want to hang out during
1:44:35 the
1:44:36 day um it's at 11: a.m. mountain
1:44:41 time so 1: p.m.
1:44:44 Eastern 10:00 a.m.
1:44:48 Pacific um and I don't there's nothing
1:44:50 up there right now so
1:44:54 um it'll be there in the morning just go
1:44:57 to my profile it's
1:44:59 there oh yeah then tomorrow's Friday
1:45:01 night date
1:45:03 night so oh Robert Rossy thank you for
1:45:06 the dyno I love that one that was
1:45:08 beautiful very generous thank
1:45:10 you
1:45:13 um so you can hang out there we can do
1:45:15 Friday night Daye night also um Saturday
1:45:18 at noon Pacific Ann Murphy does her GPT
1:45:23 weekend jams so in the AI salon and if
1:45:27 you if you're new here and don't know
1:45:29 what I'm talking
1:45:31 about go to the salon. click on the very
1:45:34 first link on that page it's a link tree
1:45:37 where it says AI Salon
1:45:39 community
1:45:41 and check it out under the technology
1:45:45 Guild is weekend jams with uh with an
1:45:49 the GPT
1:45:50 jams um so that's Saturday at noon you
1:45:53 should go to those if you're curious
1:45:55 about this [ __ ] you should go to
1:45:58 those
1:46:00 uh and what
1:46:02 else if you want to support this
1:46:05 channel you can follow me five minute
1:46:10 goodbyes you can follow the channel
1:46:12 that'll support you can subscribe to the
1:46:15 lives that supports you can buy the
1:46:18 video series that's in the corner there
1:46:19 called underhyped if you're actually
1:46:21 especially if you're in a business and
1:46:22 you're trying to figure this [ __ ] out um
1:46:25 that video series talks about why AI is
1:46:28 a bigger deal
1:46:29 than you know the people that are
1:46:32 dismissing it or demonizing it are kind
1:46:34 of missing some of the big picture and
1:46:37 so that's that's what I talk about in
1:46:38 that
1:46:39 series and the and the way that you can
1:46:42 really support this
1:46:44 channel is just keep showing up
1:46:47 here well when when when do I show up
1:46:49 Kyle I don't know show up every night I
1:46:52 do it's generally 8:00 Mountain every
1:46:55 night thank you
1:46:57 Winston I try to keep it consistent I
1:47:00 used to be just chaos I would just go on
1:47:02 any time of day and people were like
1:47:04 I'll I'll come on whenever you're on but
1:47:05 this is kind of a pain in the ass and
1:47:08 then I thought yeah [ __ ] it why don't I
1:47:10 just do it the same time every night
1:47:12 it's easy for everyone and easy for
1:47:15 me you
1:47:18 know and just show up
1:47:22 here and don't have an agenda like don't
1:47:24 even feel like
1:47:29 like I just feel like it's less and less
1:47:33 important to teach anything on this
1:47:35 channel it's called the AI learning lab
1:47:38 we want our lessons than we want them
1:47:40 now teach us
1:47:44 now I feel like the most important thing
1:47:47 is just like
1:47:49 commit to being in the conversation
1:47:53 commit to playing James come out thanks
1:47:56 thanks for your knowledge share Kyle
1:47:58 appreciate it thank you very much I
1:47:59 appreciate the kind
1:48:01 words
1:48:03 just show up here if you learn something
1:48:06 great if you show up here and what I'm
1:48:08 talking about is boring the [ __ ] out of
1:48:10 you then just hang out here and keep
1:48:12 making images or play with perplexity or
1:48:16 see if you can get it to do some coding
1:48:17 that doesn't actually
1:48:22 suck take the same
1:48:24 prompt and try it in chat GPT 3.5 chat
1:48:28 GPT 4 Claude perplexity
1:48:32 grock uh uh what else um po all the
1:48:37 different models in po like just play
1:48:40 just try different [ __ ]
1:48:42 out and go oh that was really horrible
1:48:45 but that's kind of
1:48:46 amazing Kyle tomorrow afternoon Pete and
1:48:49 Leah are going to talk about mid Journey
1:48:50 oh that's awesome Dr J yeah so just go
1:48:54 to the salon check out the events go to
1:48:56 the guilds check out the guilds they've
1:48:58 all got events going
1:49:00 on oh there's good thank you Brandon
1:49:03 there's also down at the bottom of the
1:49:05 guilds there's a thing that says
1:49:07 Community clubs there's only one
1:49:09 Community
1:49:11 Club it's this group it's called a l
1:49:16 Irregulars these are the Irregulars the
1:49:18 people that show up here all the time
1:49:20 they're the
1:49:21 Irregulars so go there scroll to the
1:49:24 bottom find Community clubs and hang out
1:49:26 in the Irregulars there's a chat area
1:49:28 where you can chat and there's another
1:49:30 area where you can share
1:49:32 images and we didn't really do any image
1:49:34 sharing tonight but I'm sure some of you
1:49:36 did anyway you can go in there and troll
1:49:38 me make funny pictures of me acting like
1:49:41 a dumb
1:49:42 Dum it's it's a great Pastime amongst
1:49:46 the
1:49:51 Irregulars and just show up just show up
1:49:54 just show up just show up just show up
1:49:56 just play play play play play and then
1:49:59 talk about
1:50:00 it why be in a
1:50:03 community so that you can talk about
1:50:05 this stuff who else should you talk with
1:50:07 this stuff about
1:50:13 strangers hey Jim how's it going it's
1:50:15 going okay hey have you played around
1:50:18 with any of this AI stuff I heard it's
1:50:21 evil let me show you something cool get
1:50:23 out your
1:50:24 phone write him a
1:50:27 song make him a picture here let me pick
1:50:31 make make a picture of your
1:50:34 Diner all right how'd you do that it's
1:50:37 AI it's cool anyway see you have a good
1:50:42 day tell your friends and family the
1:50:44 ones that tell you not to talk about AI
1:50:46 [ __ ] them tell them
1:50:49 anyway your co-workers we're not you
1:50:52 know we're not allowed to to use AI here
1:50:54 what are you why would you even talk
1:50:55 about it that seems pretty
1:50:58 risky tell
1:51:06 them Jenny's such a rebel she's just out
1:51:10 she's you know what she does on the
1:51:11 weekends she's got a Chromebook and she
1:51:14 plays with the AI she plays with the
1:51:16 chat GTP she plays with it all weekend
1:51:19 and she comes in here and all her work's
1:51:21 done she gets her work done on the
1:51:23 weekends and she comes in here and she
1:51:25 just Lolly gags
1:51:27 about be like
1:51:29 [Laughter]
1:51:36 Jenny all right so thank you thank you
1:51:39 to the Irregulars so anyone who shows up
1:51:41 here more than once you're you're
1:51:43 saintly because why would
1:51:46 you
1:51:49 um thank you to the subscribers thank
1:51:51 you to the mods thank you to anyone who
1:51:53 gifts me during the during the things I
1:51:55 don't I I feel like I don't thank you
1:51:57 folks
1:52:00 enough the mods who sort of elevate your
1:52:04 comments so that I can see them and and
1:52:06 comment to them and sort of keep me on
1:52:08 the path thank you for all of
1:52:11 that and yeah so all right let's see who
1:52:16 we talking to going to hold up my little
1:52:19 magic mirror here who do I see I see
1:52:22 Lord digital Gods I see Brandon and papa
1:52:24 JJ I see Corey Sandler Pottery you
1:52:27 should follow Cory Sandler Pottery Corey
1:52:30 c r i Sandler s n d like
1:52:34 Adam Pottery like cups and
1:52:40 [ __ ] go follow her cuz she makes awesome
1:52:44 Pottery you know what else she makes
1:52:45 awesome of music you know El she makes
1:52:48 awesome of uh stews she she does cooking
1:52:52 you know those she makes awesome of
1:52:54 [ __ ]
1:52:56 gpts she's like a GPT Master she's she's
1:52:59 a [ __ ] application developer now she
1:53:01 made a GPT last
1:53:04 week that will take an image you
1:53:06 generate in Del or anywhere frankly and
1:53:09 it will automatically add a watermark on
1:53:12 it that has a copyright With Your Name
1:53:14 and All Rights Reserved on
1:53:17 it she's an application developer now go
1:53:20 follow her anyway Brian Whitney good to
1:53:24 see you Jane Doe Christy good to see you
1:53:27 Mr
1:53:28 it show and tell uh no I'm I'm I'm
1:53:33 Crispy I just want to I just need to go
1:53:35 I just need to go
1:53:37 napping but you should go check out the
1:53:39 showand tell channel in the salon too
1:53:41 there's a whole showand tell area now we
1:53:43 have four different channels one for
1:53:45 images one for songs one for videos and
1:53:48 one for
1:53:49 gpts so go check out the show and tell
1:53:52 area Kyle show tell people are asking
1:53:54 please okay
1:53:57 fine see you you just you just can't I
1:54:00 just I wish I could quit
1:54:02 [Laughter]
1:54:08 you robot doing
1:54:12 surgery oh happy leap day oh there was
1:54:15 there was a whole bunch of cool Leap Day
1:54:17 things that one's great our road trip
1:54:19 the Dodge cult station wagon a genuine
1:54:22 lemon
1:54:23 that's flipping awesome nice Papa
1:54:27 JJ win an iPad what is
1:54:31 this oh something created using AI yeah
1:54:35 probably check out the patch on the
1:54:38 uniform yep for sure that's AI for
1:54:43 sure I need 7
1:54:47 trillion look see Cory Sandler All
1:54:49 Rights Reserved see
1:54:54 best way to spend a leap day oh that's
1:54:57 beautiful who did that cam Katen that's
1:55:00 gorgeous big
1:55:06 eyes oh that's gorgeous more leap
1:55:09 year yeah happy leap day
1:55:13 everybody class Salon oh that's gorgeous
1:55:16 Megan my head is a jungle nice
1:55:21 wow
1:55:23 yeah really inspiring I mean that's the
1:55:25 whole thing is you know just be in a
1:55:29 community with people that are doing
1:55:31 [ __ ] like this it'll it'll light you up
1:55:34 wow that's awesome I do lots of videos
1:55:36 and picks when I fly everyone says I
1:55:39 should Watermark there you go so so
1:55:42 Corey Sandler so if you go to the GPT so
1:55:45 what I'm in right now is the image and
1:55:46 Graphics show and tell if you go to the
1:55:48 gpts not not a ton have been posted here
1:55:51 so like I think two things down yeah
1:55:55 right
1:55:56 here so Corey Sandler's um art Shield
1:56:00 copyright protect watermarks that's the
1:56:02 link to her GPT you can also just go
1:56:04 look for art
1:56:05 shield and on the GPT store and it'll be
1:56:10 there I got the itch Megan got the itch
1:56:13 so get in there get in there and do it
1:56:16 all right
1:56:21 everybody let's see investor until
1:56:24 someone creates GPT how to remove her
1:56:26 Watermark
1:56:28 exactly yeah the the the watermarks are
1:56:32 just you know what it what it really is
1:56:34 is wait I'm making another one right now
1:56:37 similar but a bit different I'll let you
1:56:40 know when it's done that's great that's
1:56:41 awesome anyway go follow her Robert
1:56:43 Rossy thank you Amelio's wife good to
1:56:45 see you um Brandon thanks Danielle good
1:56:49 to see you Jane Doe
1:56:54 inhale exhale Joe 90 inhale exhale it's
1:56:57 pretty good healthy wealthy and wise
1:57:00 good handle Sherry D good to see you
1:57:02 Tobias thank you as always Daniel
1:57:04 Hollingsworth we're going to need to
1:57:07 differentiate app developers from coders
1:57:09 and Engineers if AI mashups or
1:57:11 apps I don't know that we're going to
1:57:13 need to to differentiate them Daniel I
1:57:17 think that's that's old world thinking I
1:57:19 think we are about to enter an era
1:57:23 where everyone where where all this
1:57:25 shit's
1:57:26 democratized app development is
1:57:28 democratized Art creation is
1:57:31 democratized video creation is
1:57:33 democratized music is
1:57:35 democratized all of it all of it all
1:57:38 knowledge
1:57:40 work data analysis is
1:57:43 democratized business Consulting is
1:57:45 democratized all of
1:57:48 it all the [ __ ] that we went to school
1:57:51 for where we studied a special
1:57:54 is going to be available to everyone
1:57:57 else I think the world gets really
1:58:00 interesting when that happens and we're
1:58:02 not that far from it you know we're like
1:58:04 three years from it being you know fully
1:58:06 rolled out we're we're in the middle of
1:58:08 it right now um so so yeah I think right
1:58:14 now you might want to distinguish
1:58:16 between well Corey Sandler is not really
1:58:18 a developer cuz she's just using you
1:58:20 know chat GPT that's not real devel
1:58:23 development but I can tell you that the
1:58:25 way she is developing apps a
1:58:30 developer will never have gotten there
1:58:34 cuz she's coming at it as an
1:58:36 artist with the ability to
1:58:41 code and then I look at someone like
1:58:43 Roger Scott who was a coder for 30 years
1:58:46 and the way he's approaching
1:58:49 images an artist would never get to
1:58:54 and he's making amazing art right like
1:58:55 so it's just so I think the the mashups
1:58:59 the the the societal mashups that we're
1:59:03 about to experience are going to be
1:59:04 absolutely insane and I don't think we
1:59:06 need to distinguish them I don't think
1:59:08 it's important or relevant to say this
1:59:10 one's a real developer and that one's
1:59:11 not this one's a real artist and that
1:59:14 one's not no anyone who decides to
1:59:16 express thems in any form is fair
1:59:20 game if it's good it's good how they got
1:59:24 there who gives a [ __ ] so anyway I'll
1:59:27 leave you with that tonight peace out
1:59:29 everybody keep showing up here hang out
1:59:31 at the
1:59:33 salon don't be an [ __ ]
1:59:37 peace