1/7/2025 - Unlocking AI's Power: Deep Research with Gemini

Description

Chapters

Transcript

0:01 we're doing this
0:19 [Music]
0:34 all right all right all right good
0:39 evening good evening name's Kyle
0:43 Shannon this here's your AI learning lab
0:48 and we've got what
0:49 could nearly be be described as a lab
0:53 you want to come sing champy
0:55 [Music]
1:12 came like
1:14 [Music]
1:20 slow his was wed and I look in her
1:28 eyes what my fingers yeah ever any of
1:31 you guitar players out there you ever
1:33 play guitar and like your fingers are
1:36 just like half a string off they like
1:38 you think you're hitting the right chord
1:40 but your fingers aren't
1:43 what is that
1:46 just is that is that position or is that
1:49 like my brain synaps is going
1:57 [Music]
2:08 she came on to him like slow moving
2:15 confront his beard was warmer than a
2:18 look in her
2:21 [Music]
2:23 eyes she sat on a stool and he said what
2:28 do you want
2:32 she said give me a look that don't a
2:35 love that don't try up
2:39 [Music]
2:39 [Applause]
2:40 [Music]
2:44 inside you see me pretty good tonight
2:47 you don't seem like you're all that into
2:49 it though what do you
2:51 [Music]
2:57 think every time I see you
3:00 now get that look in my
3:04 eyes every time I see your mouth
3:08 I that
3:11 smile in the early Misty morning L I
3:16 heard the
3:18 engine in the
3:24 [Music]
3:27 old you leave it me
3:31 again
3:33 today you will convince
3:36 me again
3:39 today leaving this hotel looking for
3:44 someone else's
3:48 [Music]
3:51 R so
3:56 [Music]
3:58 longan now don't you
4:04 C so
4:05 [Music]
4:10 long Don't You Cry for
4:20 Me sharing the cigarettes and keeping
4:23 warm out on the
4:26 road chasing down a lifestyle out on
4:29 Highway
4:32 24 New York state was a rolling Breeze
4:36 in a sunshine with a blue sky
4:39 falling the CH of old
4:43 September creeping
4:48 in you leave the
4:51 May again
4:55 today you convince
4:58 me again today
5:03 the shirt is no bueno oh yeah it's white
5:05 I forgot to change it into a black
5:06 T-shirt all right oh you know what we
5:08 can do we can solve
5:10 that look at that solved bang done oh
5:15 Champ's leaving he's had it he's like
5:17 that was mediocre at
5:25 best all right we need this a little
5:28 lower all all right
5:31 people good
5:34 evening my name is Kyle Shannon this AI
5:37 learning lab well we're gonna learn some
5:39 [ __ ] tonight or not I might just talk
5:42 about
5:44 nonsense um it's possible that what I
5:48 say is profound and
5:50 prophetic it is possible that what I say
5:53 is horse driil
5:57 [Music]
6:04 and I don't mind a little horse drible
6:06 out every now and again you just put it
6:08 in a shot glass and shoot it it's like
6:10 eating an
6:11 oyster you're not eating an oyster to
6:14 chew on it and savor it you just
6:19 like same with horse drible just right
6:22 down the Shute little
6:24 salty got the taste of Oats
6:29 [Music]
6:37 oh let me oh there's a lot of comments
6:39 over here hang on let me go look over
6:41 here good evening good evening good
6:43 evening all
6:45 [Music]
6:51 right Vicky I went to bed last night
6:53 with AI learning Lab videos on rotation
6:56 woke up ready to yell at the neighbor's
6:58 dog and realized it was champy
7:01 you get him champy champy will haunt you
7:04 in your dreams he will wake you up out
7:06 of your
7:08 dreams that's pretty awesome oh man I
7:13 want I love to watch big Salon TV even
7:16 with some fun Funtime
7:21 [Music]
7:26 delay okay so for everyone over there on
7:28 the on the YouTube
7:33 [Music]
7:35 um what we're going to need you to
7:38 do we hit a thousand which thank you I
7:41 like I don't we I don't know how we did
7:44 it we went from like 660 to like 720 it
7:48 took like three weeks and then I guess
7:51 maybe AI Festivus had something to do
7:52 with it but I don't think we had a ton
7:54 of people watching anyway we broke a
7:57 th000 subscribers on the YouTube channel
8:00 now we need you to watch videos and
8:03 again I'm not suggesting you watch the
8:07 videos I mean sure you
8:11 could but if you just hit play so we get
8:14 our average watch time up that'd be
8:16 swell
8:18 too is is is he advocating for fraud
8:22 that that that that doesn't seem like
8:24 very cathe I well I I mean listen I
8:28 appreciate I come here there's
8:30 occasional humor well not I wouldn't
8:33 call it humor but I mean sometimes he
8:36 thinks he's funny uh and and and and he
8:39 seems genuine about his his passion for
8:41 but I it seems to me that he's
8:43 advocating for fraud he's he's asking us
8:46 to watch videos but not not actually
8:48 watch them so so that his his his view
8:50 time numbers go up and it just seems to
8:53 me that what what he what he should why
8:57 would he not just ask us to watch them
9:00 is he
9:02 unscrupulous what I I don't understand
9:05 this channel it it's called the AI
9:08 learning lab I I I don't think I've
9:09 learned
9:11 anything I honestly I I don't think he's
9:14 qualified and and now he's asking us to
9:17 lie for him I I I just I well anyway
9:28 [Music]
10:08 does he know he's live does he know he's
10:10 live because he doesn't seem to be
10:11 looking at the
10:12 camera why he is that is that a song I
10:15 don't think that's a song it's just he
10:17 seems to be just playing strings on on a
10:21 [Music]
10:26 guitar when when when does The Learning
10:29 begin
10:40 Jim Gaffigan owes you royalties for his
10:42 comedic style which he clearly stole
10:44 from
10:46 you think it might be the other way
10:52 [Music]
10:53 around it is totally a Jim Gaffigan
10:56 thing W that's special wow wow wow which
10:59 he still from Dana Carvey that's [ __ ]
11:02 church lady Jim Gaffigan stole church
11:04 lady from Dana Carvey and I stole
11:08 Midwestern diddies from Jim
11:11 [Music]
11:24 Gaffigan you know what's all mine he he
11:26 Marge Marge Marge he
11:30 how do you know if the Tik tok's on it's
11:34 it's the one on the
11:38 phone it's I got there's there's
11:41 Hearts do do I see my
11:44 face I I see it twice I see I see if if
11:48 there's
11:50 two Mar I know you're watching the
11:53 wheel is it a good
11:57 clue if the if the hearts are on then
12:00 then I'm good there's Hearts okay okay
12:04 that's all me I mean and that's comedy
12:07 that three people have laughed at that
12:09 over the past year at
12:11 [Music]
12:28 least I'm a midwesterner now two moved
12:31 to St Louis
12:33 recently we poured midwesterners are
12:36 always picked
12:37 on is this place I can rest my PO
12:42 [Music]
12:46 head together my thoughts in sweet
12:54 silence buest place where the feelings
12:57 are de
12:59 [Music]
13:02 R an over exposure to
13:06 valence is this place I can slowly face
13:11 the only one I truly can
13:16 know these are tears from a long time
13:20 ago got these tears from a long time ago
13:25 I need to cry 30 years or so
13:30 these are tears from a long
13:35 time
13:36 [Music]
13:39 [Applause]
13:42 go all
13:45 right what are we gonna talk about oh I
13:47 got something fun so so I finally got
13:52 over being pissy about Google okay so
13:55 here's a here's a it's it's Ty raade
13:57 Tuesday so I'll go off on Google a
14:00 little bit here hopefully P's here um
14:02 because he can't do anything about this
14:03 but he does get defensive which is
14:05 pretty sweet
14:07 [Music]
14:10 um so for two years Google has been
14:13 putting out AI stuff for the first year
14:16 it was bad it was embarrassingly bad
14:18 you're like you guys [ __ ] invented
14:19 this it's it's got to be better than
14:21 this right in the past year the stuff
14:24 they've been putting out's been really
14:26 good their product strategy
14:30 is an absolute [ __ ] disaster they've
14:33 got URLs all over the place some things
14:36 are subsets of other things some things
14:38 are Standalone things some it's just
14:40 they're all over the place they' got
14:41 this notebook LM thing which was really
14:43 just a Labs experiment that went well it
14:46 was just a Labs experiment no one cared
14:48 about it and then they stuck the podcast
14:51 maker thingy in there and all of a
14:52 sudden everyone's like oh not book's the
14:54 best thing ever and they think all it is
14:56 for is for making podcasts but it's
14:58 actually really good at any that not
15:04 withstanding finding just finding the
15:07 tools on Google abject nightmare it like
15:11 worse than me trying to find a chord I
15:13 don't
15:14 know get that it's a bad guitar playing
15:17 joke
15:20 um
15:22 then my company
15:25 storyvine we are
15:27 a um Google Business customer we pay
15:32 them a lot of money every month to have
15:34 our emails and our Google Drive and our
15:37 Google Docs and our all that sort of
15:40 stuff as part of our how we run our
15:43 company and whenever they launch some
15:46 new
15:47 technology guess who's not allowed to
15:50 use
15:52 it Google workg group customers yes yes
15:56 because we pay them we're Penal
16:00 ized so I'm bitter about it but I'm a
16:04 gen exert I'm just I'm pissy about
16:06 everything right like why wouldn't you
16:08 be like you know everyone today is like
16:11 oh we want to take care of your feelings
16:13 who the [ __ ] took care of my feelings
16:15 [Music]
16:19 huh hey Kyle maybe you could just go out
16:21 maybe you could leave the house for how
16:24 long till dark you know that kind of
16:27 [ __ ] um and then when it come back in
16:30 why are you dirty well cuz I was out for
16:33 [ __ ] 12
16:37 hours so I'm bitter I've got a I've got
16:40 a a a low L it's like you you know when
16:44 you're getting sick and you don't have a
16:46 fever but you don't not have a fever
16:48 that's sort of where rage is with all of
16:50 us gen xers it's just right there it's
16:53 just right there you just have to do
16:55 something stupid like have your paying
16:57 customers not get access to the cool new
17:00 [ __ ] right and then we'll just [ __ ]
17:03 lose our [ __ ]
17:05 and right that's I stole have from Louis
17:09 [Laughter]
17:13 black
17:15 okay so today I was like oh [ __ ] it I'll
17:19 just go
17:21 pay I'll just go to my individual Google
17:24 account which is not part of work Group
17:25 which I never I don't use anymore CU
17:28 it's I had my company for 13 years so I
17:31 don't use this account and I'll just pay
17:34 this stupid 20 bucks a month so because
17:36 I wanted to go try deep research Gemini
17:40 deep research
17:41 or Gemini 1.5 Pro with deep
17:46 [Music]
17:49 research uh which I did I paid for so we
17:52 can go play with that tonight so I think
17:53 that'll be kind of fun it's it's
17:55 surprisingly good it's surprisingly good
17:58 the fact that it is is so buried and you
18:01 can't [ __ ] find it and you if you're
18:04 a work group customer you can't buy it
18:06 it's
18:08 staggering and it's possible that I can
18:10 buy it as a work group customer but I
18:12 have to reach my out to my sis admin and
18:14 it's
18:15 [Music]
18:21 just and then so that so we'll play with
18:23 that that's pretty good
18:26 [Music]
18:29 uh Jensen hang today at CES he's the CEO
18:33 he's the hip leather jacket
18:37 badass badass leather jacket wearing
18:40 Jensen
18:41 hang of Nvidia oh good lord what did
18:47 they're they're like who's gonna catch
18:51 them who's G to catch
18:54 them they're killing it on consumer
18:57 products they're killing it on
18:59 Enterprise Products they're killing it
19:00 on just [ __ ] crazy ass supercomputer
19:05 products I don't think anyone's going to
19:07 catch them I I can't believe how fast
19:10 they're moving one of the ticktockers
19:12 today said he said this is not Steve
19:15 Jobs oh it was uh what's his name who's
19:17 the guy we like Brandon uh with the
19:20 beard the dude the dude with the
19:25 beard
19:27 anyway uh
19:32 um he said it's not like Steve Jobs one
19:35 more thing he's like it was 20 more
19:39 things no not wasc no no no the he's a
19:42 tick tocker he's Nate Nate someone
19:46 Nate he talks about
19:49 AI Nate yeah Nate
19:53 exactly anyway Nate Nate talked about
19:56 the the Nvidia thing today
20:00 Nate B yeah
20:03 [Music]
20:06 exactly yeah I admire him he's he's
20:09 prolific he puts out a lot of content
20:12 and it's it's good what's what's cool
20:15 about Nate
20:18 [Music]
20:22 is I he doesn't really produce a video
20:25 until he's internalized whatever the
20:27 issue is so when he talks it's just it's
20:30 very natural I like him a lot anyway
20:33 Nate Nate B Jones that's the
20:37 dude what do you think about the letter
20:39 Sam posted about the lawsuit from his
20:41 sister I don't know family family [ __ ]
20:45 is
20:47 uh always uncomfortable I don't know I
20:52 mean I suppose it was the only thing
20:54 they could really do was do you know the
20:56 whole family against her but I don't
20:58 know
20:59 [Music]
21:00 I'll let that play out in the court of
21:02 law not for me to not for me to
21:04 [Music]
21:13 judge um
21:20 [Music]
21:42 Nvidia seems to be
21:44 releasing petabyte level Hardware it's
21:48 it's
21:49 insane their desktop did you see their
21:52 their desktop
21:53 supercomputer check this thing out wait
21:56 I gotta I gotta share different
21:59 uh stop sharing that add this do this do
22:05 this do this do this do
22:11 that everybody who uses computers today
22:15 as a tool will need a AI
22:19 supercomputer and so I just wish I just
22:22 wish that djx1 was smaller and um
22:30 effects so um
22:34 imagine ladies and gentlemen
22:38 are look at that
22:40 [Applause]
22:44 thing this is nvidia's latest AI
22:48 supercomputer so it's the size of an
22:51 iMac it can run natively it can run
22:55 their entire AI Tech stack or software
22:58 stack
23:00 um you can run a 200 billion parameter
23:04 model on it it's three
23:07 grand um it's got a Blackwell chip in it
23:10 it's three grand it's it's it's a super
23:12 it's an AI super computer in in the palm
23:15 of your hand so you can get two of them
23:17 and you can run like llama 405 billion
23:20 parameter model like the the the big
23:22 daddy models you can run on on two of
23:25 those so for six grand you daisy chain
23:27 two of those [ __ ] together and you're
23:31 running all that stuff
23:33 locally
23:35 so so there's that then they also
23:37 released they're not giving up on Gamers
23:39 they released the new GeForce 5000
23:42 series the cheapest one is
23:49 4090 uh speed for 549 bucks oops sorry
23:54 black bar black bar
23:57 people oh black
23:59 Jam back
24:03 J wait so you're on YouTube and X but
24:07 not on Tik Tok anymore no I'm still on
24:09 Tik Tok I'm on Tik Tok right now hi Tik
24:12 Tok what's shaking everybody yeah no I'm
24:15 still there I'm still there I'm still
24:19 there what's this this is NVIDIA yes
24:21 that's
24:22 [Music]
24:27 Nvidia and then they also released this
24:30 whole Suite of Robotics software
24:35 including
24:37 a world simulation model that that's
24:41 based on 20 million hours of video of
24:44 like people walking and objects falling
24:47 and whatever that
24:49 [ __ ] um plus digital twins so so
24:52 basically you're going to be able to
24:53 train up your humanoid
24:55 robots on all their software but it's
24:59 only going to run on their Hardware
25:00 right and then they released a robot
25:02 chip that's 20 times more powerful than
25:05 the last robot chip they did which was
25:07 like last year they're they're
25:09 destroying it they are destroying
25:12 it what's up AI fam hey John Harris was
25:15 shaking
25:17 [Music]
25:41 what's your favorite rock band wow
25:43 that's a great
25:45 question
25:47 [Music]
25:52 H as a high schooler it was
25:57 Boston as a guy in my 30s It Was
26:01 [Music]
26:16 Nirvana I liked RM I like Talking
26:20 Heads old school [ __ ] now it's old
26:24 school [ __ ]
26:30 [Music]
26:31 oh that's funny John John Harris did you
26:34 see did you see that chat GP got some
26:38 got some exposure to the world FBI says
26:41 that the Cyber truck guy used chat GPT
26:43 to plan his
26:45 attack that's
26:50 awesome that's the true democratization
26:53 of
26:56 AI I did not see that but I think that's
26:59 [ __ ] hilarious I mean not hilarious L
27:03 yuck yuck funny
27:04 but I what what is it any PRS good PR
27:08 well maybe the letter the letter in the
27:11 lawsuit with his family probably not
27:13 good
27:15 PR us terrorist detect use chat GPT to
27:18 plan the attack probably not good PR
27:24 [Music]
27:31 you know I would not be surprised if
27:34 although I
27:35 gotta I temper this a bit with Elon Musk
27:38 being close to the centers of power but
27:41 if if the election had gone the other
27:43 way I wouldn't be surprised if they took
27:45 something like that and really blew it
27:46 up and used it as an excuse to lock down
27:49 public access to
27:52 AI but I think that I think that now
27:55 with the the open source models out
27:58 there and the stuff that China's
27:59 released recently that's open source
28:02 like the the deep deep seek
28:05 stuff I mean I think we're pretty much
28:07 at the point where the the cat be out of
28:09 the
28:10 bag unless they make you know owning a
28:13 AI super computer and installing AI
28:16 software on it illegal but again I don't
28:18 think that's going to happen with this
28:22 Administration we can't generate
28:24 questionable images but it can help with
28:25 attacks well that's not true shanea uh
28:28 you can you can Jailbreak these things
28:31 relatively
28:33 easily just go on Reddit I'm sure and
28:36 just you know look up jailbreaking
28:39 prompts I'm sure they're still playing
28:41 with
28:43 those at this point it makes absolutely
28:45 no difference how bad open AI is chat
28:47 GPT or or PR is US Government we don't
28:52 have the capability to monitor chat GPT
28:54 usage yet
28:57 exactly but like okay so here's a reason
29:01 that that you know this little computer
29:04 here on
29:05 screen that thing can run models that
29:08 are close to if not Superior to chat
29:12 GPT and not touch the internet right
29:16 this is where we're headed where even
29:18 our phones are going to have incredibly
29:20 powerful models on them not from Apple
29:23 mind
29:24 [Laughter]
29:26 you I don't how the [ __ ] are they not
29:30 killing it in the AI
29:36 game they got the whole ecosystem they
29:38 could they could [ __ ] lock that [ __ ]
29:41 down and they're [ __ ] seeding it to
29:43 Nvidia and to Google it's
29:46 [Music]
29:55 unbelievable cam katkin sorry I'm late
29:58 all right we'll forgive you this time
29:59 cam we'll forgive you this
30:03 [Music]
30:09 time as Sam said once we know it can be
30:12 done it's pretty simple to do it again
30:14 and again apple is either missing the
30:17 boat or thinks they've got something
30:18 really big Brewing I just I don't think
30:20 they've got something really big Brewing
30:22 I just don't like like their current
30:26 Apple intelligence wouldn't suck so bad
30:28 if they had good
30:31 like you know a company like Claude a
30:35 company like I mean chat jpt is super
30:40 funded
30:41 and but they they've had Talent bleed
30:43 like there there's no [ __ ] reason
30:46 Apple shouldn't be destroying it now now
30:48 admittedly apple is always a lagard
30:51 right they're not a leader they're a
30:52 leader with if in category
30:55 definition but with [ __ ] like this they
30:57 let a lot of other people make the
30:59 mistakes and then they come in and do it
31:00 really well but I thought they would
31:02 come in and do it really well in the
31:04 fall and they didn't it just
31:06 sucks I blame Tim Apple
31:13 [Music]
31:24 exactly yeah I've been waiting for Apple
31:27 for 10 years too ever since job they're
31:29 just not
31:30 there with all that's happening one
31:32 might ask are you ready for ai.com oh
31:35 look at Brandon coming in with the solid
31:38 are you ready for ai.com so if you don't
31:41 know a couple of weeks back now week and
31:43 a half we did this thing called AI
31:45 Festivus it was it was two 12-h hour
31:48 days on a Friday and Saturday it was a
31:50 work shopathon called AI Festivus and a
31:53 lot of people were saying could we turn
31:54 this into a training and we're in the
31:56 process of doing that right now so if
31:58 you want to pre-register for that it's R
32:01 you readyfor ai.com
32:05 ready for AI get it the number four
32:08 instead of f r are you ready for ai.com
32:17 um and
32:19 uh yeah our Target launch date is
32:22 February 1st
32:24 [Music]
32:30 Apple intelligence equals chat GPT
32:33 1.35 pretty much it's pretty [ __ ]
32:36 [Music]
32:37 shitty all right let's go play with
32:41 Google deep
32:42 research since I [ __ ] paid for it AI
32:45 festiv this was awesome have you all
32:48 talked about Nvidia 32 gigabyte vram
32:51 going to be released at end end of
32:53 January need
32:54 to oh you need it for deep seek well we
32:57 we talked about this thing um I I there
33:02 was so much that Nvidia talked about
33:04 today there there's this
33:06 machine there's the 5000 series of the
33:09 GeForce GPU cards there's all that
33:12 robotics stuff is just crazy
33:15 like Nvidia is going to single-handedly
33:18 enable humanoid robots to to be launched
33:20 in 2025 I have a
33:22 feeling I I mean they're going to be
33:24 launched anyway but Nvidia is going to
33:26 make them really good
33:29 um all right so here we
33:32 are I this is oh wait oh because I was
33:35 sharing sound hang on a sec God dang it
33:39 you
33:41 frakin I I want to know what engineer
33:44 decided this is how screen sharing was
33:46 going to work and I want to take them
33:48 into the back alley and have my way with
33:51 them well that that sounded lwd
34:01 uh cam Ken exciting I'm a fan of
34:04 Robotics and AI yeah I I'm telling you
34:06 man the the um the accessibility stuff
34:09 that these robots are G to have you seen
34:11 that one that one [ __ ] robot out of
34:14 China you you think you're scared
34:18 now let's see uh let's see
34:23 China robot
34:26 Wheels I think that should be all we
34:28 have to do no not that one this
34:32 one wait let let me let's get the audio
34:34 for this one this one's pretty [ __ ]
34:39 up it's like yeah we're gonna be fine
34:43 everything's gonna be fine nothing to
34:46 nothing to be scared of AI ai's ai's oh
34:52 okay that's fine
35:01 look at
35:02 [Music]
35:10 that crazy there's another one where
35:12 it's blasting down a
35:14 hill is that the one let's see this
35:18 [Music]
35:25 one woke up today feeling brand new
35:27 nothing in my way so going to sh yes
35:30 here to stay and I'm going to say what I
35:32 need to say hey I'm feeling free
35:35 everybody watch doing me go ahead look
35:38 you going see that I'm exactly where I'm
35:40 supposed to be I'm doing me yeah just
35:44 doing
35:46 meing yeah
35:49 I'm I'm
35:55 doing I'm
35:57 [Music]
36:08 no wonder they want to ban Tik
36:10 Tock Tik tok's the least of our
36:14 worri 2,000 of these in the woods behind
36:17 your house is is what's next on the on
36:20 the Worry
36:24 List oh good lord anyway let's go do
36:27 some research um okay so here's what
36:30 we'll do with the with deep research so
36:32 if you don't
36:34 know uh tabs I got to change my tab back
36:38 right yes of course because we got to
36:40 share tabs because we can't we couldn't
36:43 possibly just share once and have it
36:45 work no
36:49 that you know what it is the software
36:52 Engineers have decided to give us gen
36:54 xers an excuse to be pissed off that's
36:56 what it is okay all right so here we are
37:01 at this
37:02 thing ticktock
37:05 pin Mars Rover upgrade yeah to say the
37:08 least I mean did you see that thing like
37:11 crawling up the the waterfall coming
37:13 down on those smooth rocks have you ever
37:15 tried to walk on smooth rocks like that
37:17 they're slipperier than
37:19 [ __ ]
37:21 um all right so here's what we'll do
37:24 since we got some good activity in the
37:25 chats what do you want me to research so
37:28 so let let me let me walk you through
37:31 where where we are here we're we're at
37:35 um
37:36 gemini.com
37:38 and I subscribed to the advanced
37:43 advanced
37:44 Gemini with my individual account you
37:47 can't do this with a work group account
37:50 because
37:52 I I
37:53 can't but and then now that you have
37:57 this you've got five different models
38:00 you can choose from so I AI is getting a
38:03 little bit better at this they're
38:04 they're kind of dynamically assigning
38:06 models but like look at the look at the
38:09 explanation of 2.0 flash
38:13 experimental here here's when should you
38:15 use 2.0 flash experimental might not
38:18 work as
38:21 expected okay but what does it
38:24 do poly Z says hi on Tik Tok hey poly
38:28 what's
38:28 happening what is
38:31 happening
38:33 [Music]
38:39 um so anyway they've got this one of the
38:42 models in here is called 1.5 Pro with
38:45 deep research and what it does is what
38:47 it sounds like it will research
38:49 something so who wants me to research
38:53 something going to need three of those
38:55 some costume Prosthetics and a vacation
38:58 to the
38:59 Appalachians research AI memory issues
39:02 how can we kind of work around it what
39:05 do you mean AI memory issues do you mean
39:06 like context window being too
39:08 short ASI I like the memory feature on
39:12 GPT but it's always full
39:15 ah any other AIS that have more memory
39:19 uh I don't think so right now but it's
39:24 possible I am very wrong on
39:26 that I don't think the Gemini models
39:29 have memory right now um open ai's
39:33 Memory went from 8,000 tokens to 32,000
39:36 tokens all of them all of them are
39:39 talking about in the labs right now
39:42 they've got million token memory sizes
39:45 and they think they're going to be able
39:46 to get to infinite memory which when
39:49 they get to infinite memory the the how
39:51 we use these tools is going to
39:52 completely alter I think I think part of
39:55 the real answer is just wait
40:00 my llm remembers 100% of everything now
40:02 Joker's just showing off because Joker
40:04 built a lang chain if if you want memory
40:08 go build your own go use Lang chain
40:10 create a an embeddings
40:12 database and and go build yourself some
40:15 something with memory yeah really good
40:18 research Ai and Healthcare testing
40:20 medication on tissue samples to find the
40:22 best treatment individualized medicine
40:24 okay so here we are we're in this I wish
40:28 I could turn the brightness down on this
40:29 monitor I probably can but that's too
40:32 [ __ ] much sorry ticktockers that's
40:34 just the way it is uh let's see can I
40:38 does any of this help no all right
40:41 whatever um
40:43 okay
40:47 um reearch
40:52 oh and then I've got this issue are you
40:54 kidding
40:56 me somehow I've got my um some setting
41:01 somewhere on my
41:08 Mac Dark theme ask
41:12 Gemini okay we'll do it in light theme
41:16 uh some some settings somewhere on my
41:19 Mac someone is not coding their
41:21 applications right for my machine
41:25 because I don't I I don't think other
41:27 people have problem but I just had I
41:29 just had white text on a light gray
41:31 background uh box solve for lime Kyle
41:34 yeah no [ __ ] but let's do we'll we'll do
41:36 something a little more broad uh to to
41:38 test this so we'll go we'll go
41:42 um research the
41:46 very latest in
41:54 individualized medicine on a
41:59 C
42:01 Kate gory
42:05 by
42:06 category
42:09 basis basis
42:13 and
42:20 um
42:22 focus on advancments
42:31 driven by
42:34 generative Ai and Deep
42:38 Mind
42:43 Etc I would also like to
42:52 know what stage the medicines are in
43:00 and
43:02 estimated
43:04 time to
43:06 Market I don't know we'll just start
43:08 there let's see let's see how that
43:11 does all right you can tell chat jbt to
43:14 modify summarize or completely remodel
43:18 your hang on your memory by the way oh
43:22 that's
43:23 cool I didn't know that I knew you could
43:26 go in there and you could manage
43:27 manually I didn't realize you could tell
43:29 it to update [ __ ] because it's it's
43:32 really random what it decides to
43:34 memorize it always memorizes the weird
43:35 [ __ ] for
43:40 me oh oh okay so so here we go okay so
43:44 the way this works is I I asked I asked
43:47 it
43:48 to research all that [ __ ] it says you
43:51 know here's basically what I wrote so
43:53 now it's going to give me its plan find
43:56 categories of indiv indiv ided medicine
43:59 for each category find the latest
44:01 developments for each category in one
44:05 find the latest developments within each
44:07 category focus on advancements driven by
44:09 generative Ai and companies like deep
44:11 mind if there are too many such
44:13 advancements found limit to several that
44:15 are most representative for each
44:18 advancement find the stage of
44:19 development for each advancement find
44:21 the estimated time to Market if
44:23 available find any challenges or
44:26 limitations okay and let's let's so what
44:28 I'm going to do is I'm going to go edit
44:30 the plan and this is weird interface I
44:32 say edit the plan it retypes the plan
44:34 this is not editable but I can just
44:37 prompt it which you can prompt it
44:38 without saying edit the plan but that's
44:40 all right uh I'm gonna say
44:44 um
44:46 also add
44:51 country
44:53 uh
44:55 research is being
44:58 conducted and how
45:02 conservative
45:05 their uh
45:07 FDA
45:09 type
45:11 organizations
45:15 are Kyle on the AI newsomatic it it says
45:18 last updated 55 years ago yeah I know
45:21 cuz I I ran out of well it'll as as soon
45:25 as it updates when you say update it'll
45:26 it'll go from 55 years to you know two
45:29 seconds ago it's just a temporary state
45:31 where it says 55 years um basically I
45:35 got that thing to the point I was ready
45:37 to launch it and then I completely ran
45:39 ran out of monthly credits at lovable so
45:42 I could probably go back now and fix it
45:44 and do another pass on it um but I just
45:46 ran out of [ __ ] so I ran out of tokens
45:49 and I didn't feel like paying more money
45:51 because I'm like that's kind of a minor
45:53 a minor bug but yeah thanks for pointing
45:55 that out I know that's there C GPT
45:57 memory hack copy it into a document and
46:00 then create a custom GPT and use the
46:02 document as a knowledge document update
46:04 is needed there's even there's even a
46:06 cooler one in can you do this
46:11 in I don't remember if chat GPT now
46:14 reads Google Docs I don't think it does
46:17 but I think je uh Claude does so you
46:20 could do you can do kind of a custom GPT
46:22 over in Claude anywhere where you
46:26 can uh do it on as a Google doc um
46:29 whenever you update the Google Doc it'll
46:31 just see it you could also do that with
46:33 um notebook LM
46:36 okay find the categories of
46:39 individualized medicine for each
46:40 category final okay let's see what it
46:42 does here at the bottom for each
46:44 advancement estimated time to Market
46:45 finding the challenges for each
46:47 advancement find the country where the
46:49 research is being conducted for each
46:50 country find out how conservative their
46:52 FDA type organizations are that's fine I
46:55 guess that's fine all right so now we're
46:56 going to say start research boom and so
46:59 this thing's going to go off now and do
47:01 it and it's kind of fascinating to watch
47:04 it it you know grabs websites so
47:07 genome.gov two Wikipedia things a Penn
47:11 State thing a video at Penn State um
47:16 FDA so it started with five we websites
47:20 now there's
47:22 76 so it's going and finding all these
47:25 websites and articles and blog posts and
47:27 [ __ ] like
47:28 that now it's up to 231 so this is going
47:32 to 231 websites and this number may go
47:36 up like this we we gave it a pretty
47:38 meaty thing this is going to take a
47:40 while but like imagine just going to 231
47:46 websites much less reading them and
47:49 comprehending them in any way navigating
47:51 to the thing you need to get to this is
47:53 incredible isn't it John it's
47:56 incredible um
47:58 all right what's it doing over here so
47:59 it's researching so here's your little
48:01 here's your little timer right so so you
48:03 got your little progress bar so it's not
48:05 even close right now it's just
48:07 researching so I guess it's out reading
48:09 all these websites look at
48:13 this can we click on any of these do any
48:15 of these take us there yeah they do yeah
48:18 so like while it's researching you could
48:20 go look at the sources it's finding stem
48:22 cell
48:24 therapy at the data Intelligence Center
48:30 cool are we still at 231 yeah
48:33 researching 231
48:37 websites Bonkers isn't
48:41 it Global crisper genan
48:44 editing lot of crisper lot of crisper
48:46 going on
48:49 here very cool if you can please post
48:51 the results in the salon oh that's a
48:53 great idea okay I'll do that um I oh the
48:57 other thing that this does again I Pate
49:00 are you here I'm saying oh yeah you are
49:02 you have 231 tabs open right now so I
49:05 don't see the big deal but the
49:08 difference is this thing actually reads
49:10 those 231 tabs Pate I never go back to
49:13 them um I just want P to know I'm saying
49:16 something good about Google so one of
49:18 the cool things that this does is it
49:20 goes and you know researches all these
49:23 websites it then writes up a report I
49:26 can then give it notes like if I don't
49:28 like the report it gave like I did one
49:30 this afternoon where I wanted it to do
49:32 research on AI video personas in healthc
49:37 care marketing and what it did was AI in
49:40 health care I'm like no I don't want
49:42 that so I had to redo it a bunch of
49:43 times so you can redo the
49:45 report and then once you're happy with
49:47 the report you basically just say open
49:50 this as a Google doc so once this does
49:52 what it does we'll we'll make it better
49:55 now now it's analyzing the results see
49:57 our little circle here's you know at
49:59 8:30 8:30
50:02 o00 um but yeah I'll be able to I'll be
50:05 able to share this as a uh as a Google
50:07 doc I'll just make it available for
50:09 everyone
50:10 yep it would take me 231 hours minimum I
50:14 know if you've got add you go to one of
50:16 those sites and then spin off like I
50:20 like I could I couldn't get to 20 sites
50:23 much less
50:25 230 now we're at 9:00
50:28 individualize analyzing results still
50:30 analyzing so I guess 231 was our final
50:32 number of research
50:44 sites oh I can probably show
50:48 you yeah let's
50:53 see AI personas so here's here's what I
50:57 did AI generated video person personas
51:00 Revol revolutionizing Health Care in
51:03 marketing the rise of AI in healthcare
51:06 and Pharma AI generated video person
51:09 personas a new frontier in
51:12 communication so it sort of T like
51:14 what's cool about what this wrote was
51:17 this gave my sales guy a bunch of
51:20 talking points about how to talk about
51:22 this stuff because it went and found
51:23 just all the articles that said anything
51:25 about the kind of thing that we're
51:28 selling and then it ethical
51:30 considerations key players so there's
51:33 the there's the vendors here are
51:36 individuals people that are into it
51:39 really pretty slick
51:43 stuff all right where are we now what's
51:45 going on oh we're
51:47 done advancements in individualized
51:50 medicine a detailed report so how long
51:53 is this looks like it's pretty
51:55 long pharmac Cod
51:58 genomics
51:59 so when I was at Penn State studying
52:03 acting I did I did a well I don't have a
52:06 degree but I did a a second um I guess
52:09 you'd call it a weekend study program on
52:12 pharmacogenomics because no I
52:18 didn't what is it it's the study of how
52:20 genes affect a person's response to
52:21 drugs oh that's cool I get it C genomics
52:25 offers a critical pathway for Tailoring
52:27 drugs and therapies oh that's cool so if
52:29 you can figure out if you can figure out
52:31 from someone's genes and and DNA how
52:35 they respond to drugs then you can okay
52:38 the use of clustered regularly
52:39 interspaced short palindromic you know
52:42 this is really stupid research for me to
52:44 do because I can't even [ __ ] read the
52:49 responses AI driven
52:52 advancements Alpha fold all right cancer
52:56 treatment G editing individualized
52:59 cancer treatment for transforming wait
53:01 individualized medicine is transforming
53:03 cancer treatment Gene editing
53:05 crisper um immunotherapy so it looks
53:08 like it's giving us an overview first
53:10 rare
53:11 diseases oh you know what I'm going to
53:12 do I am G to throw in so here's what I'm
53:14 gonna do I'm gonna go um I want you to
53:23 throw throw in God get you
53:28 throw in lime
53:34 disease it's
53:36 co-
53:39 infections and
53:42 oops Li disease
53:45 ah typing in public is hard people you
53:48 think this is easy it's co
53:55 infections and
53:59 related L hul
54:03 issues boom all right so we're gonna
54:06 have it go off and do that I I think it
54:08 just does that in the background now
54:11 deep Minds alpha alpha
54:18 missense here's my plan
54:22 oh find the latest an individualized
54:26 okay da da da fine yeah go do that start
54:31 research oh it's going to it's it it ate
54:34 our thing all right well we'll wait for
54:35 it we we're gonna we're gonna add some
54:38 some personal Research into this larger
54:42 document found 21
54:47 websites if you don't know we've got
54:51 chronically ill twin boys upstairs right
54:54 now dealing with this shitty ass disease
54:57 all right you're right my ADHD would not
54:59 let me get past the first six sites hash
55:02 take my
55:07 money next thing you know you're gonna
55:09 be making a song onso instead of
55:10 researching I'll totally go
55:16 there so it's another another instant
55:19 blog Creator yeah this is this is well I
55:23 I mean here's what I don't know about
55:25 deep research I don't know I don't know
55:29 how
55:31 um how well their anti- hallucination
55:35 checking is
55:38 um I assume because it's an llm it's got
55:41 all the
55:42 hallucinatory goodies that all of the
55:44 other ones do like I know I know
55:46 perplexity is pretty bad at
55:47 hallucinating so I don't know if this
55:49 one's any better but the fact that this
55:50 is going to so many websites is just
55:54 staggering so I found 39 websites for
55:57 Lyme disease good now it's doing its
56:01 thing all
56:03 right used it to get the
56:05 state-of-the-art survey paper of the
56:07 current performance optimizations for
56:09 inference llms oh that's cool yeah it's
56:13 like I mean if you think
56:16 about if okay here here's a good way to
56:19 put a tool like this into
56:23 perspective if you if you think that for
56:27 the past two years generative AI has
56:29 been put in the hands of non-engineers
56:31 right so people are out here doing all
56:34 this work and you know some of us are
56:35 just playing here and [ __ ] around and
56:38 playing with AI but there's also
56:40 researchers out there and people trying
56:42 to solve problems that have never had
56:44 access to AI that all of a sudden have
56:45 access to these tools so for two years
56:47 they've been using these tools Deep Mind
56:49 comes out and their material stuff comes
56:51 out and Alpha fold and Alpha fold 2
56:53 Alpha fold 3 all that stuff's coming out
56:56 all that stuff going out into the
56:58 world it's going into Labs they're doing
57:01 stuff with it they're writing articles
57:02 about it they're writing peer-reviewed
57:04 papers on
57:05 it and it is happening at an incredible
57:09 rate and so what this thing does
57:12 assuming it's doing its search well it's
57:14 Google it better [ __ ] do its search
57:16 well it's going out and finding all the
57:20 stuff
57:21 that would be hard to find and hard to
57:23 keep up with and it's doing it like that
57:25 so
57:27 just
57:28 amazing all right what are we doing okay
57:30 we're still we're analyzing the results
57:32 we got 39
57:34 sites huc hallucinations aside anyone
57:37 would be researching for facts at some
57:39 point after this so I'll take it exactly
57:41 I mean that's with with with all of
57:44 these tools with every AI tool whether
57:46 it's an image generating tool a song
57:47 making tool chat GPT this thing when it
57:51 generates what it generates use your
57:54 thinky brain
57:56 use critical thinking look at it
57:59 research it does you
58:01 know spot check some of the references
58:04 right if it's got a citation click on
58:07 the citation go look at the thing say
58:10 does this say what that said and did
58:12 this actually exist that's one of the
58:14 things perplexity does is it will invent
58:17 citations so you see citations you're
58:19 like oh these are all great but like
58:21 three of them might be not
58:23 non-existent so I assume this one's
58:26 better at that part of it just subscribe
58:28 to Gemini thanks Kyle now I have another
58:29 AI subscription I know John I'm sorry
58:32 about that like like I said I raged
58:34 about this for what a month Pate
58:37 bitching about I can't play with any of
58:39 the new Google [ __ ] because I'm a paid
58:42 paying
58:43 customer all right wait now it's just
58:46 about Lyme
58:47 disease H wait do I still have the other
58:50 one I think I do
58:56 yeah oh okay cool so this is just like
59:00 um this is just like uh what's it called
59:07 um chat GPT C canvas I don't think I can
59:10 go in here and edit this but these
59:13 little thumbnails on the left are are
59:15 these different articles so I can open
59:17 this in Google
59:18 Docs and it formats them really nice in
59:21 Google
59:23 Docs there it is whoop d is all right
59:27 you can all see that right and then
59:29 because it's Google Docs it's got this
59:31 nice table of
59:33 contents all right phenomic far
59:36 pharmacogenomics cancer treatment rare
59:38 diseases aiid driven advancement in rare
59:41 diseases neurological
59:44 diseases AI advancements and
59:46 neurological diseases stage of time to
59:49 Market cardiovascular
59:52 diseases and then there's little oh the
59:55 the citation are not hyperlink that
59:58 seems like a
1:00:00 Miss come on
1:00:02 people make it a
1:00:06 hyperlink all right here's a table okay
1:00:09 despite the significant progress in
1:00:12 individualized medicine challenges
1:00:13 remain in terms of cost comple
1:00:16 complexity and Regulatory
1:00:18 hurdles
1:00:21 advancements AI driven advancements
1:00:27 then work cited so it's a relatively
1:00:28 short research thing but just even
1:00:31 getting the list of work cited so you
1:00:33 can go look on your own and this will
1:00:36 give you a starting point so what I'll
1:00:37 do is let me share this so I'm going to
1:00:41 make this available to hang on I got to
1:00:43 do a stupid [ __ ] dexterity Challenge
1:00:46 on Tik Tok uh anyone with the link I'm
1:00:49 gonna copy the link then I'm going to go
1:00:51 over to the AI
1:00:53 sell where am I going to put this I will
1:00:56 put this in llm showand
1:01:02 tell if you don't know what the AI Salon
1:01:05 is shame on
1:01:08 you go there the salon.
1:01:12 a click on the thing that says join our
1:01:15 community and this is that Community
1:01:17 it's a remarkable Community we just had
1:01:19 a a meeting tonight and AI Salon
1:01:21 presents
1:01:22 tonight um the salon is also what
1:01:26 co-sponsored the um AI Festivus that
1:01:30 everyone loves so much all
1:01:33 right
1:01:35 why is this
1:01:40 [Music]
1:01:45 Hung okay so when you're in AI Salon
1:01:49 there's a little show inel area so I'm
1:01:51 going to go to llms and gpts I'm going
1:01:53 to say here is the
1:01:58 AI
1:02:00 advancements
1:02:01 in healthc care dock I created
1:02:07 with
1:02:08 Gemini
1:02:12 1
1:02:14 five Pro with
1:02:19 deep
1:02:21 [Music]
1:02:24 research paste
1:02:34 don't notify and then I think I will go
1:02:37 back to this here thing let me go look
1:02:40 at the I actually know something about
1:02:42 lime disease you know what's shitty when
1:02:45 you learn some a lot about something and
1:02:47 you didn't want to know about
1:02:49 it think I really wanted to know about
1:02:52 Lyme disease
1:02:54 no I could write a book or to I should
1:02:58 all
1:02:59 right advancements and
1:03:03 challenges diagnostic tool we don't care
1:03:05 about those too too late for diagnosis
1:03:09 we're we're solid on that black bar yes
1:03:12 what you said
1:03:14 Brandon what is better Gemini 2 or the
1:03:17 new 1.5 little confused there Joker I
1:03:21 don't Paul rer said this on his on his
1:03:24 podcast this week he's like
1:03:26 all of these all of these research
1:03:30 companies need to
1:03:32 stop giving us a choice of models how
1:03:36 the [ __ ] are we supposed to know what is
1:03:38 the difference between 40 40 mini 0101
1:03:43 mini and
1:03:46 four unless you're a geek like me that
1:03:49 pays attention to this [ __ ] but I don't
1:03:51 know the difference between all the
1:03:52 Google
1:03:54 models as Paul rer said
1:03:58 today they're creating the most
1:04:01 sophisticated knowledge tools on the
1:04:04 planet couldn't they also use that same
1:04:07 tool to say look at your prompt and
1:04:10 identify which of the models it should
1:04:13 use they could do that again it would it
1:04:19 would require a product strategy Google
1:04:21 clearly doesn't have open AI clearly
1:04:24 doesn't have they're a research lab
1:04:26 Claude doesn't have I mean they're
1:04:27 trying they're bringing on product
1:04:29 people but they're not you know they're
1:04:31 not consumer companies
1:04:32 by you know well Google is Google Google
1:04:37 should know better but they fired all
1:04:38 their marketers it's all Engineers now
1:04:41 all
1:04:42 right the clim health
1:04:45 app all right individual stage of time
1:04:48 to Market challenges and limitations
1:04:53 conclusions while challenges Remain the
1:04:56 advancements discussed here offer hope
1:04:58 for improved outcomes for future line
1:05:01 patients the AI diagnostic
1:05:03 tool
1:05:05 synthesis a few years not
1:05:09 specified
1:05:12 yeah it's
1:05:18 okay see like I think what I would do
1:05:21 here oh that's I'm not in the thing
1:05:23 anymore think what I would do here is I
1:05:25 would go
1:05:28 um
1:05:30 identify the
1:05:34 lime
1:05:37 coinfections and follow
1:05:43 on
1:05:45 conditions and research them all
1:05:51 individually and then give me a consult
1:05:57 validated
1:06:01 report it's going to come up with this
1:06:03 little
1:06:04 plan
1:06:15 sh John John Harris is already doing
1:06:18 some deep research over there okay John
1:06:21 and everyone else that's listening
1:06:23 so why do I do this why do I do this
1:06:26 night after night after night after
1:06:28 night do you think it's for the pink
1:06:32 bow just so I have an excuse to wear a
1:06:34 pink
1:06:36 bow you're correct
1:06:39 and what John just did here he's like
1:06:43 watching me he's like oh God is he ever
1:06:45 gonna stop singing is he ever GNA do
1:06:46 anything and then and then I stumble on
1:06:49 this thing and he's like oh [ __ ] there's
1:06:52 one of those while we're here he goes
1:06:55 and subscrib
1:06:57 and now he's doing this the whole point
1:06:59 of this channel is just to be in the
1:07:02 conversation be chatting with people be
1:07:04 saying hey have you tried this have you
1:07:06 tried that there's like all sorts of
1:07:08 coaching going on on the back channel
1:07:09 here it's really cool to watch that's
1:07:12 exactly why we're
1:07:15 here I already have too many screenshots
1:07:17 of you had the pink bow at this point I
1:07:20 am unblack mailable with the pink bow
1:07:26 all right oh start research let's get
1:07:28 this thing
1:07:29 going all
1:07:36 right Source Camp Kelly Camp thinks I
1:07:39 really am here for the pink bow listen I
1:07:42 didn't I did not even know about pink
1:07:45 bows pink bow hair thingies and and by
1:07:48 the way I wear it wrong for all you men
1:07:51 out there next time you're trying to
1:07:53 impress your lady as the sensitive guy
1:07:56 type don't put it on like this they'll
1:07:59 make fun of you see you got to pull it
1:08:02 down way low in the back and then you go
1:08:05 hi I don't see I don't know how to do it
1:08:09 but
1:08:09 you see because when I put on my
1:08:13 Foundation you don't want your hair in
1:08:15 your face
1:08:17 right you know a little blush Little
1:08:19 Foundation powder it up that you get the
1:08:22 hair out of your face that's how you
1:08:23 wear
1:08:24 it not like not like you
1:08:28 know Ivon lendel on a you know on a
1:08:32 [Laughter]
1:08:37 dare Andre agasse on a dare when he
1:08:40 still had hair this is more Andre Agy
1:08:43 than Ivon lendel Ivon lindel was all
1:08:48 uptight see there's a screenshot I
1:08:50 didn't have me wearing the pink bow
1:08:53 correctly so many people from f of us
1:08:56 messaged me that the pink B well about
1:08:58 the pink bow character not knowing it
1:09:00 was
1:09:03 Kyle oh wait not KN yeah yeah is that
1:09:06 right not knowing it was Kyle yeah
1:09:08 that's
1:09:13 awesome oh you
1:09:15 know we have fun well I have fun I don't
1:09:20 know I don't know what you people are
1:09:21 doing here so many people from Festivus
1:09:24 that's really funny
1:09:26 oh
1:09:31 man John Harris having yet another Kevin
1:09:34 McAllister moment there you go that's
1:09:37 why we do
1:09:38 this if if you don't know what a Kevin
1:09:40 McAllister moment is remember the movie
1:09:42 Home
1:09:43 Alone that guy um if you use AI long
1:09:48 enough and it basically here's the deal
1:09:50 if you just stay in the
1:09:52 conversation show up to [ __ ] like this
1:09:55 join community
1:09:58 be unreasonable about how much time you
1:10:00 spend like looking at what's
1:10:03 coming um at some point you're G to just
1:10:07 stumble on another thing you're gonna go
1:10:09 holy [ __ ] holy [ __ ] there's another one
1:10:12 there's another one there's another one
1:10:13 happens all the
1:10:15 time for me
1:10:19 so and every one of those little moments
1:10:21 where you have this Epiphany like I
1:10:23 didn't even know that was possible you
1:10:25 now know that something new is possible
1:10:28 and the next time you go to solve a
1:10:30 problem you go oh I could use that deep
1:10:33 research thing for this right that's how
1:10:35 this works right now there's no better
1:10:38 way to do
1:10:39 this certainly not with ad but I think
1:10:42 even even for methodical people for you
1:10:46 ocers let's stick with the neuros spices
1:10:49 for you ocers out there it's not any
1:10:51 better because you can't keep up there's
1:10:53 just too much volume to do it
1:10:55 methodically
1:10:57 you can't just do it randomly right so
1:11:00 the the only way I know is to be like a
1:11:02 shark moving all the
1:11:04 time and just play just play with that
1:11:06 toy play with that toy play with that
1:11:08 toy Chomp on that one oh that one's
1:11:10 really good let's eat all of that one
1:11:12 let let's dig down on that one and then
1:11:14 just back back into the waters just keep
1:11:16 it moving keep it
1:11:17 moving only way I
1:11:19 know had so much fun teaching another
1:11:22 experimental class at tww they were
1:11:24 blown away oh that's so much fun Kelly
1:11:27 Texas Women's University right what did
1:11:29 you show them what were your what were
1:11:30 your parlor tricks and let me know if
1:11:32 you want to come up here and talk about
1:11:33 it I'm happy to have you up um have you
1:11:36 seen projects in chat GPT what is this
1:11:39 just popped up projects is cool I can I
1:11:41 can show that stuff um oh this is nice
1:11:46 look here we go babesia we've got those
1:11:51 uh anaplasmosis I don't think we have
1:11:53 that ER liia we might have some of that
1:11:55 that Bartonella we got Bartonella we're
1:11:58 like a we're like a Bartonella pickling
1:12:00 Factory Here
1:12:03 bellia Maya maty I don't know that one
1:12:06 all right but
1:12:08 anyway all
1:12:10 right anything more better down
1:12:22 here Diagnostics drug and development
1:12:25 advancement and stem cell and
1:12:27 nanotechnology
1:12:29 challenges challenges money no one's
1:12:32 paying for this
1:12:34 [ __ ] all right cool I'm gonna open that
1:12:36 in
1:12:41 docs topic was finding a Target client
1:12:44 but showed voice and voice with video
1:12:48 yeah that one that that one's a a
1:12:49 mindblower the uh the voice and voice
1:12:52 with video
1:12:56 all right let me let me pull out of this
1:13:06 thing oh you wanted to see projects okay
1:13:10 hang
1:13:11 on let me show you
1:13:15 projects so projects is is very similar
1:13:19 to projects in
1:13:22 Claude um so over here on the left
1:13:25 you've got these project things now
1:13:28 let's
1:13:29 see there's a stupid thing so see I've
1:13:33 got this project called
1:13:34 Sydney so here's Sydney and I've only
1:13:37 got one chat in it but I've done like
1:13:39 dozens of chats about this
1:13:41 musical so if I go in here and I type
1:13:44 Sydney it's gon to find all my chats and
1:13:47 unless they' fixed this which I'm sure
1:13:49 they haven't there is no way for me to
1:13:51 get one of these chats into a project
1:13:57 because they're [ __ ] if I share it
1:14:00 there's nothing there
1:14:03 no so what someone told me is that oh
1:14:07 sorry what someone told me is that what
1:14:10 you can do is you can say
1:14:16 hello and it'll
1:14:19 respond and
1:14:22 now let's see let me hit new chat and
1:14:27 let me see
1:14:35 reload
1:14:41 reload AI projects brainstorming no it
1:14:44 didn't do
1:14:46 it
1:14:49 h well
1:14:51 anyway projects is cool so long as you
1:14:54 don't have to search for past
1:14:55 conversations but let me let me tell you
1:14:57 what what projects
1:14:59 are um so when you've got a project you
1:15:02 can add files to it you can add custom
1:15:05 instructions to it so it's like it's
1:15:07 like a custom GPT in that sense right so
1:15:11 you can tell it you want it to act like
1:15:13 a screenwriter or whatever you want it
1:15:15 to act like you can add files so you
1:15:17 could so actually what I could do is I
1:15:20 could go find all my old chats and I
1:15:23 could add them as files but you can also
1:15:26 add
1:15:27 chats right so you can take any chat
1:15:30 that's over
1:15:31 here and if you click on this three
1:15:34 dotted line you can say add it to the
1:15:38 project okay so this is again why why
1:15:42 this you know the the the developers
1:15:45 they just want to keep us gen xers
1:15:47 pissed off
1:15:49 Sydney so they've already
1:15:52 got they've already got a a programming
1:15:54 object
1:15:56 to take a list of
1:15:58 chats and have three dots to the right
1:16:01 that do something and they chose not to
1:16:04 [ __ ] put it in this
1:16:09 interface right these three
1:16:12 dots they could add this they they've
1:16:15 got this object they have this
1:16:18 object and and no one at at the the $700
1:16:23 trillion open AI with the big [ __ ]
1:16:27 bigheaded
1:16:28 Geniuses no one did a search for old
1:16:31 chats and was like hey uh what if we
1:16:34 wanted to add one of these search
1:16:36 results into a project no one no one no
1:16:40 one really anyway so sorry John I was
1:16:44 telling you about projects but once you
1:16:46 get all your chats in here you can now
1:16:49 chat with all of your Collective chats
1:16:52 like this thing will now know all of the
1:16:54 files you added it'll act like your
1:16:56 custom instructions and it'll tap into
1:16:58 all those previous chats you had so it's
1:17:00 a really [ __ ] slick little thing with
1:17:03 with one major flaw like if you've been
1:17:05 using chat GPT for two years wouldn't it
1:17:08 be swell if you could organize all those
1:17:10 chats into a
1:17:13 project and listen I hope one of you
1:17:17 points out my stupidity and says duh
1:17:20 Kyle you add things from the search
1:17:22 results by dragging and dropping them or
1:17:24 whatever this stupid [ __ ] thing is I
1:17:26 would love that but I haven't found it
1:17:29 [Laughter]
1:17:33 yet that was your chat Kyle it popped up
1:17:37 what was my
1:17:38 chat I just imported my history to
1:17:42 obsidian yeah I know but that like it's
1:17:45 like they've got the data right here
1:17:47 they have an object to be able to add
1:17:49 things to a project just put that in the
1:17:52 [ __ ] search
1:17:53 results did does open AI need to hire
1:17:59 me like
1:18:02 seriously I'll just be like the bitter
1:18:04 guy in the corner they like they bring
1:18:06 over a new launch they we're thinking of
1:18:08 launching this to the public all right
1:18:11 here why can't I search on on my past
1:18:15 chats well you can in the mobile app
1:18:18 yeah but I use the website all day
1:18:27 oh wait my most recent chat
1:18:30 is AI brainstorming
1:18:37 project oh yeah oh it is all right fine
1:18:40 all right fine but that still pollutes
1:18:43 my that still pollutes my my chat
1:18:47 history anyway it I'm let me be bitter
1:18:51 let me be
1:18:52 bitter fine that thing worked you can
1:18:56 you can search for a thing okay so if I
1:18:58 search for Sydney I can go Sydney
1:19:00 musical rewrite and then I can go oh see
1:19:03 hello wait no it's
1:19:06 Sydney musical
1:19:09 rewrite that's not my latest
1:19:14 one all right now I'm totally [ __ ]
1:19:17 confused Sydney musical rewrite hello
1:19:20 how can I assist you
1:19:21 today um by not
1:19:28 sucking fair
1:19:33 enough all right so now let's let's go
1:19:35 new chat and then let's go
1:19:39 reload Sydney musical rewrite okay there
1:19:41 so it's there so I can now add this to I
1:19:44 can now add this to the project add to
1:19:48 project but do you have do you see do
1:19:50 you see where I'm right and everyone
1:19:53 else is wrong
1:19:54 [Laughter]
1:19:57 by not
1:20:00 sucking anyway they should hire me and
1:20:04 they they they should give me a corner
1:20:06 office and I should just do nothing but
1:20:09 like sip whiskey and do [ __ ] nothing
1:20:13 and then about the time they're like
1:20:15 yeah we think we got something here I'll
1:20:17 look at it and within five minutes go uh
1:20:21 why can't this [ __ ] do that and
1:20:24 they'll be like well
1:20:28 logically and I'll just go no go fix
1:20:36 it good Lord do you give courses on what
1:20:42 comedy
1:20:43 rage rage product development yes I'm
1:20:47 doing a workshop next week um I will do
1:20:50 courses I I have done courses uh I just
1:20:53 did AI Festivus with an Murphy we did
1:20:55 did we did uh 24 sessions in 36 hours so
1:21:00 yes yes lots of courses so feel free to
1:21:03 DM me if you want to do want to talk
1:21:05 about that I'm happy to talk about it um
1:21:08 I'm just silly here you know we're just
1:21:10 silly we're just silly you know what we
1:21:11 do you can make money we're just silly
1:21:15 here we'll just little birds we're just
1:21:17 Little Birds trying to make software
1:21:19 [ __ ] work
1:21:29 oh just a matter of time before they
1:21:31 Scout you for the job Kyle it's gonna be
1:21:33 a long
1:21:36 time I think they should hire
1:21:38 [Laughter]
1:21:44 Frank oh good lord all right people it's
1:21:48 it's late I started a little late but
1:21:50 let's get let's get y'all out of here
1:21:56 oh good good good good good good this
1:22:00 was fun this fun we have fun we have
1:22:05 fun all right
1:22:08 um
1:22:12 okay go check out do I have this written
1:22:15 anywhere I don't think I do oh maybe I
1:22:17 have a can you put up a banner Brandon
1:22:19 for are you ready for ai.com um
1:22:25 so Brandon's gonna put up a banner right
1:22:27 now Daisy May amazing as always yeah I
1:22:30 mean amazing
1:22:31 amazing I I thank you thank you I I mean
1:22:35 I appreciate the kind words
1:22:40 amazing I I
1:22:42 mean enigmatic
1:22:45 unique interesting
1:22:53 interesting Rachel Sparkles agrees with
1:22:56 Daisy May it's a cult it's a
1:23:01 cult okay on screen right now are you
1:23:04 ready for ai.com so that is the course
1:23:08 that Anne Murphy and Vicky and I are
1:23:11 putting together um based on kind of the
1:23:15 the insights that were shared during AI
1:23:18 festivas we're turning it into a
1:23:20 learning program a
1:23:22 training um and we're we're calling it
1:23:26 an AI Readiness
1:23:28 training and I'm I'm so excited about
1:23:32 that name because and about that concept
1:23:35 because I used to talk in the AI Salon
1:23:38 about AI
1:23:40 literacy and I think we're in a we're in
1:23:42 a
1:23:43 weird
1:23:45 surreal surreal place right now
1:23:50 where the the concept of being AI
1:23:53 literate is actually a
1:23:56 fallacy like there are no experts right
1:23:59 now no one is AI literate everyone's
1:24:02 trying to figure it out everyone's
1:24:03 trying to figure it out so being AI
1:24:06 literate I don't think you can actually
1:24:08 achieve right now things are changing
1:24:09 too fast we we're we're in an
1:24:12 accelerating parabolic or a uh uh Whata
1:24:15 call it um uh exponential
1:24:20 curve oh thank you very much for the
1:24:22 gifts Omar appreciate that
1:24:25 um oh my view on Tik Tok is way off oh
1:24:28 that's unfortunate isn't
1:24:30 it that's better look I'm pretty am I
1:24:33 pretty am I pretty
1:24:35 mama you have nice hair okay
1:24:42 um what was I talking about uh oh it
1:24:46 doesn't [ __ ]
1:24:50 matter your face was blocking your face
1:24:55 thank you for whoever sent me gifts on
1:24:57 Tik Tok I appreciate oh thank you very
1:24:59 much Omar Silver Fox thank you so much
1:25:01 that's that's very very sweet of you uh
1:25:03 Omar title gifter oh yeah [ __ ]
1:25:06 knocking them down thank you very very
1:25:09 much I appreciate that um oh are you
1:25:12 ready for AI so so what this this idea
1:25:15 of AI
1:25:17 Readiness I think you
1:25:19 can train
1:25:21 yourself to be ready for AI and what
1:25:24 that starts to to look like is you start
1:25:27 to get more Curious you start to get
1:25:29 more adventurous you start to get more
1:25:30 flexible and adaptable you start using
1:25:33 your critical thinking like instead of
1:25:36 just taking what AI gives you as a as
1:25:39 good look at it and if it's shitty tell
1:25:42 it it's shitty right like all that sort
1:25:44 of stuff um thinking intentionally about
1:25:47 what problem you're trying to solve
1:25:49 first and then get to the tools not how
1:25:53 do I make money with you can make Mone
1:25:56 with the reason I make fun of these
1:25:59 people is this is backwards you can make
1:26:02 Mone with chat GPT no you can't make
1:26:06 money with chat GPT any more than you
1:26:08 can't make money with
1:26:10 Photoshop you make money solving a
1:26:14 problem if you happen to use chat GPT to
1:26:17 solve that problem well then you've made
1:26:19 money with chat gbt but it's solving the
1:26:21 problem is the thing people pay for
1:26:23 right so just getting your head around
1:26:26 how you live in an insane world where
1:26:30 the technology capabilities keep going
1:26:32 up and up and up and
1:26:34 up and we don't know if the skills that
1:26:36 we've developed for prompting and prompt
1:26:38 engineering and all the [ __ ] we do
1:26:40 today is even going to be relevant three
1:26:42 months from now probably will may not
1:26:45 be so why you would do something like
1:26:48 the the AI Readiness training that we're
1:26:51 putting together is
1:26:55 just to [ __ ] change your mindset
1:26:57 change your
1:26:58 mindset and I think it also includes
1:27:00 things
1:27:03 like and this is this is going to be
1:27:05 depressing for a lot of people initially
1:27:07 but it I I find it gets really exciting
1:27:10 is you kind of have to let go of how you
1:27:13 currently value your contribution to the
1:27:18 world because you might be really really
1:27:21 good at data entry and you love data
1:27:24 entry and data entry is the
1:27:27 best
1:27:28 well that job's going to be automated
1:27:31 out but there's something about how you
1:27:33 do what you do and the knowledge you
1:27:35 have and the expertise you have that is
1:27:38 where the real value is it's not in the
1:27:40 data entry it's how you do your job it's
1:27:42 how you live your life it's the
1:27:43 Integrity you
1:27:45 have and that's a shift that I think all
1:27:47 of us are going to have to make and
1:27:49 really ask the question what do I want
1:27:50 to do what do I want to contribute
1:27:52 what's my real value here
1:27:55 and then we lead with that and then we
1:27:57 figure out what tools to use so that's
1:27:59 that's what this thing's all
1:28:01 about all right look a squirrel yeah I
1:28:05 know well what are you gonna
1:28:07 do welcome to chat add Daisy may I feel
1:28:11 incredibly fortunate to have found you a
1:28:12 year ago at least I'm aware of what's
1:28:15 happening and and that's that's all you
1:28:18 need to be Daisy may like you're like
1:28:20 paying attention you're in the
1:28:22 conversation because mark my word like
1:28:25 like the the Deep research thing we did
1:28:28 tonight like that's something that at
1:28:30 some point someone's going to come here
1:28:32 or I'm GNA come here one night and go
1:28:35 okay here's the thing we've all been
1:28:37 waiting for it's here it is truly
1:28:39 magical even Pate can't deny that it's
1:28:42 got magic in
1:28:44 it and it's GNA change
1:28:47 everything that's probably not going to
1:28:49 happen like on a single night but it
1:28:51 could it might
1:28:55 but the only know only way you're going
1:28:57 to know that that moment comes is to be
1:28:59 in the conversation whether it's with me
1:29:01 or some other
1:29:03 community hard to get my head around
1:29:05 that Kyle I know I know I know this
1:29:09 is I don't think we as a species have
1:29:13 ever had to deal with the the amount of
1:29:15 change we're about to deal with I mean
1:29:20 short of like cataclysmic [ __ ] you know
1:29:23 where just everything changes all at
1:29:25 once it's so funny I I might have talked
1:29:27 about this on here one
1:29:29 night I almost feel like
1:29:32 Co was preparation for what's coming
1:29:36 with with AI because what Co taught us
1:29:40 was we as a global Community can just
1:29:44 change the
1:29:45 rules we can just say don't come in to
1:29:49 work
1:29:50 anymore and the whole world doesn't now
1:29:53 it takes something remarkable to shift
1:29:56 to make that happen but it happened so
1:30:00 we now know that we now know that that's
1:30:03 possible so AI has has a very real
1:30:06 possibility of having some kind of
1:30:08 impact like that is it going to happen
1:30:10 overnight I doubt
1:30:12 it but if it even if it happens slowly
1:30:16 over five years in the big scheme of
1:30:19 things that is remarkably fast for
1:30:21 things to change dramatically and it's
1:30:24 Poss
1:30:25 if you look at Sam alman's letter where
1:30:27 he talks about you know we're going to
1:30:29 have ai workers in the workforce in
1:30:32 202 some [ __ ] might be changing
1:30:36 so start thinking about yourself all
1:30:40 right
1:30:42 is2g I will turn this AI car around
1:30:45 right now if I hear someone say
1:30:48 magic pate's not having it okay math
1:30:51 magical it's it's math magical all right
1:30:54 I'm getting out of here people hope you
1:30:56 had fun tonight um it was chaotic it was
1:30:59 zany but I think we did some cool [ __ ]
1:31:03 um yeah if you want to do the research
1:31:05 then go play with that Google deep
1:31:07 research it's it's pretty swell all
1:31:09 right so peace out on YouTube LinkedIn
1:31:12 and X I will see you tomorrow's
1:31:14 Wednesday it'll be normal time 8:00 pm
1:31:17 tomorrow um go check out are you
1:31:19 readyfor ai.com sign up uh early if if
1:31:22 uh if you're interested in that John
1:31:24 Harris thanks Kyle that was fun thank
1:31:26 you so much peace out to
1:31:29 the to the web
1:31:32 folks there