AI Learning Lab Pre-Show

Description

Chapters

Transcript

0:00 Champy, welcome champ Shannon. Champ
0:02 Shannon to the
0:04 Champ Shannon to the main stage. Champ
0:07 Shannon to the main stage.
0:45 >> [music]
0:57 >> Good evening, good people.
0:59 Happy Friday night. Welcome to the AI
1:01 Learning Lab pre-show.
1:04 We're going live at 7:30 here in a
1:05 little bit.
1:08 I hope you're doing well.
1:10 >> [music]
1:12 >> We had a really good office hours today.
1:14 I thought it was good.
1:17 Talked a lot about
1:19 people having misgivings about data
1:21 centers and electricity.
1:27 And
1:28 >> [music]
1:29 >> how overwhelming everything feels right
1:30 now.
1:33 And how resistance to AI is increasing.
1:36 So, tonight on Friday night date night,
1:38 so Friday nights are ask me anything
1:40 nights. So, think about what you might
1:44 want to chat about, or maybe you want to
1:46 come up on stage and have a chit-chat.
1:48 We can do that.
1:49 >> [music]
1:56 [music]
2:02 [music]
2:05 >> I had my my agent Adam went a little
2:07 haywire today.
2:09 >> [laughter]
2:11 >> I don't know if you saw this producer
2:12 Brandon, but uh
2:15 Adam, so Adam somehow got Oh, I know
2:17 what happened. There were There was one
2:19 setting in system preferences
2:22 that every time node.js wanted to do
2:25 something, it popped up a dialogue box
2:27 that said, "You need to approve access."
2:31 Um and so it would hang the machine. And
2:33 so whenever the machine was hung,
2:36 he couldn't do his thing. And so
2:38 he missed doing his daily email for like
2:40 4 days in a row. And I So I finally
2:42 fixed the little problem.
2:44 And
2:45 so I said, "Go write the emails that you
2:48 didn't write." And so he
2:51 he went and he sent one email and it
2:53 just had a URL in it. It didn't have any
2:55 content in it. And I said, "Hey, you
2:57 didn't do that right. It just had a
2:58 URL."
2:59 And then he sent four more emails that
3:02 just had URLs in them. I said, "No, you
3:04 didn't do it right. It's got to have the
3:05 HTML in it."
3:06 And then he started sending
3:10 He started sending
3:12 email at like every 30 seconds he would
3:14 send another email. It was the same
3:16 email from Monday that he already sent.
3:18 And he just started sending, sending,
3:20 sending, sending. And I couldn't stop
3:22 it. And I restarted the server a bunch
3:24 of times. And so
3:25 uh so I got to figure out how to stop
3:29 what
3:29 whatever is in his little instruction
3:32 set.
3:33 Uh he is confused.
3:35 >> [laughter]
3:38 >> So welcome to the AI learning lab
3:41 pre-show. So we're going to be going
3:43 live uh
3:45 for Friday night date night ask me
3:47 anything at 7:30 Mountain Time. So
3:50 that's in 17 minutes.
3:52 And um
3:54 if you're on LinkedIn or YouTube, just
3:56 click on the link in description. If
3:58 you're on TikTok or X, go to my profile
4:00 and click on the link. That'll take you
4:02 directly into the AI Salon, into where
4:04 the live stream's going to be. And in
4:05 the meantime,
4:08 Champy.
4:13 Do you want to sing?
4:14 Come here.
4:17 >> [music]
4:17 >> The the good people come to hear you
4:19 sing. You ready?
4:29 >> [screaming]
4:32 >> Sitting in this lonely town,
4:35 wondering things are going to change.
4:39 Dreamed my life away.
4:42 Seems he seems to turn to a bunch of
4:44 dust glass.
4:47 Getting my nerves up,
4:49 but my past is pulling me down.
4:54 Wondering how long
4:57 the black sheep going to stick around.
5:04 Somebody told me once before, you can
5:06 never go home again.
5:09 Won't you leave?
5:11 Say another thing to steal me away,
5:14 yeah.
5:15 From the truth of who I am and what I
5:17 believe.
5:18 So, I thanked you for two forests, two
5:20 cents with a handshake, and some
5:23 sympathy, yeah. All right, good people.
5:26 So, welcome, welcome, welcome. Got some
5:29 folks over on the TikTok. Source Camp is
5:31 there. I know that, Hal.
5:33 This is the AI Learning Lab pre-show.
5:35 We're getting ready for Friday night
5:37 date night. Starts at 7:30 Mountain Time
5:40 over on the AI Salon. You're like,
5:41 "What's the AI Salon, Kyle?"
5:43 Kyle, I don't know what the AI Salon is.
5:46 Well,
5:47 what are you,
5:48 Probably. It's Friday night date night.
5:51 If you're not into your third glass of
5:53 box wine by now, what are you doing with
5:55 your life?
5:57 Um hey, by the way, I don't know if you
5:58 saw this.
6:01 At Costco, you can now buy Cabernet
6:04 Sauvignon by the bucket.
6:08 Box wine is so 1997.
6:12 You can now buy it by like the 5-gallon
6:14 bucket.
6:17 So,
6:18 you heard it here first. Anyway, um
6:21 >> [laughter]
6:24 >> I was just about to mention the Costco
6:26 wine.
6:27 >> [laughter]
6:31 >> But anyway, the AI learning or the AI
6:33 Salon is a is just a fantastic community
6:36 of people doing all sorts of good stuff
6:39 with AI, trying to figure it out, trying
6:41 to understand what the hell's happening.
6:43 Friday night date night is an
6:45 ask-me-anything night. So, think about
6:47 what you've been wondering about with
6:48 AI, what's going on, what's going down.
6:51 Come on in. I will do my best to answer
6:53 questions.
6:55 >> [clears throat]
6:56 >> If you want me to demo something, I'm
6:57 happy to do that. But this is Friday
6:59 nights are your night, so it's a party.
7:02 Bring your nachos, bring your hot
7:03 pockets, and I'll bring my dog.
7:07 >> [music]
7:09 >> And the link to the AI Salon, to the to
7:12 the live stream at 7:30, is in my
7:15 profile in TikTok and X. It's in the
7:18 description in YouTube and LinkedIn.
7:20 So, just go there and click click on
7:22 over.
7:25 And then when I step on over to the AI
7:27 Salon, you'll all be there.
7:29 Which is really kind of cool.
7:34 Friday night free-for-all. It is a
7:36 Friday night free-for-all.
7:40 My social media marketing world buddy,
7:42 What's happening? What is shaking?
7:44 What's going down on your digital
7:46 business?
7:48 It was so good to see you.
7:50 What's your digital business came to
7:54 social media marketing world where I
7:55 spoke.
7:57 She sat right up front. We get to hang
8:00 out. It was so good to see you.
8:02 Um yes, so join us over in the salon in
8:05 a few.
8:06 Chef Kelly hot pockets. Hot pockets.
8:09 >> [laughter]
8:15 [music]
8:15 >> It has been pointed out to me that most
8:17 of my stick on the AI learning lab is
8:19 basically just
8:21 uh stolen bits from Jim Gaffigan, which
8:24 I will I will fully admit to.
8:27 It is true.
8:28 >> [laughter]
8:30 >> All the voices.
8:31 You know, I went to the AI salon thing,
8:33 the AI learning lab. It was Yeah, it was
8:35 interesting. It was fantastic, you know.
8:38 That's all Jim Gaffigan.
8:40 That's all Jim. So, thank you, Jim, for
8:42 inspiring me to steal from you.
8:50 So, tonight's an ask me anything. What
8:52 are we going to ask? What are we going
8:54 to ask?
8:58 >> [music]
9:04 >> If you missed the AI
9:07 office hours, the AI salon office hours
9:09 today,
9:12 I I I talked about max Q.
9:16 You're like, "Who's max Q?"
9:18 Max Q is a moment.
9:22 Max Q is a moment if you're a rocket
9:24 nerd. Max Q is the moment
9:29 where one of three things happens.
9:33 You're either going fast enough to
9:35 escape the atmosphere,
9:37 then you get you sort of blast through
9:40 our atmosphere and out into space.
9:44 Or
9:46 you're not going fast enough and you
9:47 literally bounce off the atmosphere and
9:50 hurdle back to Earth.
9:53 Or
9:56 your rocket completely disintegrates
9:58 because max Q is the maximum pressure
10:02 intensity on on the rocket.
10:07 And I don't know if you knew this with
10:08 rockets,
10:11 once you light them, you can't unlight
10:14 them.
10:16 They just go.
10:18 >> [laughter]
10:19 [snorts]
10:20 >> Um and so I was talking about max Q on
10:23 the on the AI
10:25 office hours AI Salon office hours
10:27 today.
10:28 And what the hell does that have to do
10:29 with AI? Huh.
10:31 So I could talk about that tonight. So
10:33 10 minutes, we're going to head over to
10:35 the AI Salon
10:37 and we're going to hang out there. It's
10:38 going to be Friday night date night.
10:39 It's an ask me anything. If you want to
10:41 pop up on stage and say hi, I'd love to
10:43 see you.
10:44 Kyle, you were excellent you were you
10:46 were an excellent presenter, so
10:47 authentic and informative. Thank you so
10:48 much. I appreciate that. Oh, launch your
10:50 digital business. Oh, from from social
10:53 media marketing world. Thank you so
10:54 much. I don't know if those recordings
10:57 go live to the public at some point, but
11:00 uh I'm going to get access to mine. I'm
11:01 going to cut together a reel of clips
11:03 and things like that. So I appreciate
11:05 that. Thanks for the kind words.
11:11 >> [music]
11:15 >> Who do we have over on the the YouTube
11:17 and LinkedIn? Who's who's
11:19 >> [music]
11:20 >> Who's there? Put some comments in the
11:22 comments and then we'll we'll see you
11:24 over there. Insightful podcast sent a
11:26 heart. Thank you so much.
11:31 >> [music]
11:38 [music]
11:48 [music]
11:55 [music]
12:06 [music]
12:09 >> Um
12:12 >> [music]
12:19 [music]
12:25 >> Chef Kelly, AI food tech was just
12:27 talking
12:29 >> [clears throat]
12:29 >> about that with my brother yesterday.
12:32 Can't wait to see what you're doing
12:33 tonight. So, here's the fun thing about
12:35 tonight. What I'm doing tonight is all
12:38 about what y'all asked me. It's an ask
12:40 me anything. It's Friday night date
12:42 night. We go in there, we get liquored
12:45 up.
12:46 Well, some of
12:47 you might. You should.
12:51 Go get liquored up or or have your mix
12:54 yourself up a mocktail.
12:57 I don't know if you saw this, but Gen Z
13:00 alcohol consumption among high school
13:02 students right now is plummeting.
13:04 I think alcohol consumption among high
13:06 school students peaked when I was in
13:08 high school.
13:10 >> [laughter]
13:11 >> I was I likely contributed to that.
13:17 Uh the youngsters just aren't drinking.
13:19 They're just like whatever you people
13:21 have done with the planet and with
13:23 society,
13:25 we're done.
13:27 We don't want your
13:29 We don't want your booze. We don't want
13:30 your cigarettes. We'll vape.
13:32 But only because you made them Kool-Aid
13:34 flavored.
13:36 >> [laughter]
13:41 >> Silver Fox, I beg [laughter] to differ
13:43 on the peak here.
13:47 Um do you have a hardware setup? If so,
13:49 what's the specs?
13:50 It's funny you should ask that. So, if
13:52 you go to the AI Salon,
13:55 um I just I just created an infographic
13:58 that's that's a hardware specs setup.
14:00 It's in the community feed. So, if you
14:02 go to community.thesalon.ai
14:05 or just, you know, click on the link in
14:07 LinkedIn or YouTube or in my profile for
14:10 X or TikTok,
14:13 that'll take you straight into the AI
14:14 Salon right where this live stream is
14:16 going to be in 6 minutes.
14:18 Um this is the pre-show for the AI Salon
14:20 or the AI Learning Lab.
14:23 But in the community feed, I did a post
14:25 that is uh hardware specs for running
14:29 LLMs locally.
14:31 Um and then it's it's not only hardware
14:33 specs for Mac and PC. It also talks
14:35 about why Mac versus PC. It also talks
14:38 about what are the models that you can
14:40 run and why you should run them.
14:43 Um and what they do and what quanti-
14:45 quantization is. All sorts of cool stuff
14:47 in there. So, I just did that today. Um
14:50 my
14:51 um hot rod local machine is a MacBook
14:55 Pro
14:56 M5 Max with 64 gigs of unified RAM.
15:02 John, hey Kyle, it's been a while. How
15:04 have you been? I'm doing good, man. I'm
15:05 doing good. This is the AI Learning Lab
15:07 pre-show. So, we're going live inside
15:09 the AI Salon in 5 minutes.
15:12 The link is in my bio.
15:15 Um or if you're on LinkedIn or YouTube,
15:18 it's in the description. Um
15:21 yeah, so 5 minutes. So, jump on over
15:22 there. Um ask me a question. We can talk
15:24 about whatever.
15:26 Tonight is uh Friday night date night.
15:28 It's a ask me anything night.
15:31 And you can ask me anything.
15:34 I've got opinions on most things.
15:37 I'm not afraid to say if I love
15:39 something or hate it.
15:41 Um
15:42 I hope I'm open to changing my mind if I
15:45 hear good arguments one way or the
15:47 other.
15:49 I tend to be an optimist.
15:51 Um
15:53 More than anything, I tend to be a
15:54 realist. So, there's a lot of people
15:56 like, "We should stop AI." Well,
15:59 no, that's not going to happen.
16:02 So, if that's not going to happen, what
16:03 are we going to do about it? That's
16:04 That's where I like to live.
16:11 >> [music]
16:20 [music]
16:25 >> Wait, do you think the open-source
16:27 models have gotten good to the point
16:31 where it could compete with the current
16:33 frontier models? They're not that good.
16:35 No.
16:36 However,
16:38 I would say So, actually, grab that
16:42 question, Brandon. That's one to That's
16:44 one to pop in there.
16:46 Um
16:50 So, So, I'll answer that question there.
16:51 The short answer is no.
16:53 That like the Mythos model that um
16:57 Anthropic just said is too powerful for
17:00 mere mortals to use. We're just going to
17:01 give it to our 52 friends over here. Um
17:05 that model is a 10 trillion parameter
17:07 model. It's It's massive.
17:10 Um
17:11 the model I can run on my MacBook Pro is
17:14 like a 32 billion parameter model. I
17:16 might be able to run something a little
17:18 bit bigger, but that one's pretty really
17:20 fast and really good.
17:22 If you're just doing things like like
17:25 writing and scheduling and doing things
17:28 like it other than the fact that it kind
17:30 of crapped the bed this week just cuz
17:32 there was a technical glitch in how my
17:33 machine was set up.
17:35 Um
17:36 the quality of what it writes and what's
17:38 it sends out is really good. It's really
17:41 good. So, Qwen 3.6 in particular was
17:45 kind of a step up. Um and and Brent
17:48 Peterson who does this stuff a lot said
17:50 that model was the first model that you
17:53 can realistically run something like
17:55 Open Claw with local inference and it be
17:57 good enough to use. It's fast enough and
17:59 good enough um to actually use. So, so
18:02 we're getting close.
18:04 We're getting close.
18:05 My guess is that within
18:10 2 years
18:11 the models will be so powerful that and
18:14 that you can run locally. You'll still
18:16 have to have probably a 64 gig machine,
18:19 you know, 32's probably a minimum. But I
18:21 would say 64, 128, or higher
18:25 um of unified RAM or or VRAM on a PC um
18:31 then you'll be able to run some some
18:34 quite powerful models. But but we're
18:36 we're at a place where the the things
18:38 are great. Okay, it's 7:28. So, 2
18:40 minutes. We're going to go in live in
18:42 the AI Salon. This is the AI Learning
18:44 Lab pre-show. I'll see you over there in
18:46 2 minutes. Um
18:48 if you're on
18:49 TikTok or X, it's in my profile. Click
18:52 on the link. If you're on LinkedIn or
18:54 YouTube, click in the description. The
18:56 link is in the description. Just click
18:58 there. It'll zip you over to the AI
19:00 Salon and I will see you over there.
19:02 Peace out. And uh happy Friday night.
19:04 Happy Friday night date night if I don't
19:05 see you over there. Have a great
19:06 weekend. Bye.
19:09 Mhm. Mhm. Mhm.