AI Learning Lab Pre Show

Transcript

1. 0:10Came out of the cave black as coal. His beard was warmer [singing] than a look at the fire.
2. 0:27She sat on a stool and he said, "What do you [singing] want?" She said, "Give me a [singing] love that don't freeze up inside."
3. 0:50Happy Friday night date night. Good evening, good people. We are here. First real weekend of summer. The uh F1 boys are racing in Monaco.
4. 1:06>> [music] >> So, uh if you're an AI learning lab regular, you're probably on your yacht. >> [laughter]
5. 1:15>> In the in the harbor there in Monaco. You know, Money Carlo, the No. Anyway, uh this is the uh AI learning lab
6. 1:26pre-show. It's Friday night date night. We will be going live at 7:30 Mountain Time here in the grand state of Colorado.
7. 1:36We're here in the $20 million production studio. Made to look like a shitty home office. Um if you want to know, well, you like,
8. 1:46"Where's the AI Where's the AI salon?" We're going live inside the AI salon. What's that mean? Well, if you're on the
9. 1:51YouTube or the LinkedIn, you go to the the description below. There's a hyperlink in there. If you're on the Tik Tok or the X, you go to my
10. 2:00profile. You click on the link in there. That'll take you over there to the AI Salon. That's where we'll be going live tonight.
11. 2:06Friday nights are ask me anything nights. So, if you have any questions, you can ask them here. Um and we'll producer Brandon will pick
12. 2:15them up and migrate them over and we'll answer them over in the show. There's a lot of stuff going on. Um Both uh Open AI and Anthropic are
13. 2:27threatening to launch new revolutionary models. >> [music] >> I actually don't think these these model releases are as big a deal
14. 2:42as they used to be. Cuz the models are already so good that the the improvements are likely above where you're using them on
15. 2:53a daily basis where um where it will start to become apparent is if you're using these new agentic platforms like Codex
16. 3:03thought code and things like that. [singing] >> [music] [music]
17. 3:23[music]
18. 3:43>> On Tik Tok, pin my goal so more people will know it. So, here's what Tik Tok is telling you. I have a goal that they have given me to
19. 3:54do stuff that you're supposed to do. So, do it. Engage. Engage.
20. 4:12Tick tock. This whole gamification of engagement is so silly. Dora the TikTok star. >> [music] >> She was the mistress of engagement.
21. 4:30>> [music] >> Ricky ticky Tabby's in the house. Happy Friday night date night. We'll be going live here in the AI Salon in about 15
22. 4:39minutes. Welcome, welcome, welcome. Mimi in the making is in the house. I see Silver Fox. >> [music]
23. 4:46>> Is agents a fad? I'll definitely talk about that. Um I'll definitely talk about tonight that tonight. So, so tonight's an ask me
24. 4:57anything. Um agents are not a fad. Agents have historically the the word agent has been so misused and so I've been hearing about AI agents
25. 5:14and the difference they're going to make in the world for probably 15 years. And what they used to mean was basically
26. 5:21a logic tree like like Siri and Alexa where, you know, thousands of programmers are program- programming in all the possible ways
27. 5:30that you could talk to, you know, a thing and it would do things for you. Um they called those AI agents. They
28. 5:38weren't really AI agents. Those were those were programmatic um response trees. >> [laughter] >> Now, they used some AI to do speech to
29. 5:49text and, you know, there was AI in them, but they weren't AI agents. And then, in the past probably year and a half to years, a lot
30. 6:01of people have been using the term agent for automation. They've been using it synonymous with automation. Um
31. 6:10and sometimes agents were referred to as, you know, voice agents. If you're If you're talking to something, you know, the the
32. 6:16things that Siri and Alexa wanted to be, right? Using AI just as a thing you can talk to. But, these new agents, these new agentic
33. 6:24platforms, um they're they're not a fad and and they're um I think this is where all of the frontier model companies are going.
34. 6:34I think that what starts to happen is uh they become our primary way we use AI. I think for the for the people just
35. 6:43getting into AI right now, um they're going to seem really complicated. If you wait for a year, year and a half, they're going to be
36. 6:51pretty simple. You're just going to talk to them. Hello, my friends. My late No, Mr. RT, you're right on time. If you're here,
37. 6:57we're early. This is the pre-show. So, welcome everybody to the AI Learning Lab pre-show. We're going to be going
38. 7:03live in the AI Salon. It's an ask me anything night, so think of any questions you might have. If you're on uh Tik Tok or X, go to my
39. 7:10profile. The link to the salon is there. If you're on YouTube or LinkedIn, it's down in the description. Just click
40. 7:17click on the the link down in the description. >> [music] [music]
41. 7:33[music] [music] >> Rapa papa rapa [singing] >> Mr. IT, I've created two agents using
42. 7:49open claw and claw. I'm so excited. Very cool. Here's the thing. I created two agents. That's awesome. Congrats. Um the hype cycle
43. 7:59Wait, the hype cycle probably might be then a fad. Well, AI agents as pre-AGI. Um
44. 8:16We'll we'll talk about that. As hyped as everything is, I don't really know that we've entered the hype cycle yet. I don't think we
45. 8:27I don't think we actually understand what's coming. There's there's some things that have been talked about in the past few days that uh
46. 8:38Is it Anthropic? Yeah, I think it's Anthropic. Anthropic just did this blog post that that basically we're not quite at
47. 8:48recursive self-improvement, but we're really close. And when we get to recursive self-improvement, like it's just going to be hard to I
48. 8:58can't I can't fathom what's coming. But you know, right now we're getting a new incremental model release about every 6
49. 9:07weeks. That used to be every 6 months. When we get to recursive self-improvement, it's going to be like
50. 9:14every 6 days, then every 6 hours, and just the the capability of these these tools is just going to go like that. What the
51. 9:23what these agentic platforms do is they allow you to essentially set a goal. There was an old I'll talk about this
52. 9:34when we go in there. So so anyway, um so join me at 7:30. We're going to go go live inside the AI Salon. This is the AI
53. 9:41Learning Lab pre-show. And so, if you're on LinkedIn or YouTube, in the description is the link to the AI Salon. If you're on uh Tik Tok
54. 9:49or or X, it's in my profile. Um but these uh these agentic harnesses are just going to allow us to kind of set goals.
55. 10:05Well, what do you mean a goal? Like a goal, like I want to start a business that makes me money is a goal.
56. 10:13And then, these agents will go do stuff and go figure stuff out. So, we were talking on on office hours today, we
57. 10:21were talking about this uh and there's there's a guy in there, Sam. Um and Sam went from like being like, I
58. 10:30don't know much about AI to became he became like a master prompt writer. And he got himself a job, and then in
59. 10:36his job he got promoted, and buh buh buh buh buh buh buh buh buh. And in the last month or two, he he said
60. 10:43to his work colleague, "If we don't get working to learn about Claude Code or OpenAI Codex, we're going to be left behind, so
61. 10:52we should do that." And he said, "It's absolutely transforming things because he doesn't need to prompt anymore.
62. 11:00He just needs to know what he wants, let the thing go off and give him three options. Go build me three things and bring me back
63. 11:10options." And the these things will just go off and do it. So, it it is AGI is is more about the overall
64. 11:18capability of the models, um but these agentic platforms are about the ability for each of us to be able to deploy
65. 11:27virtual employees that work independent of us. Um so, it's a it's a very different thing and there's, you know, we we
66. 11:34actually actually now finally have platforms that are truly agentic platforms. Um, Open Cloze is one of them, Codex, Cloud
67. 11:44Code, things like that. So, we'll talk about that tonight. All right, how is your app working? Mimi. >> [music] [music]
68. 11:59>> It is Friday night date night. It is indeed, Sharon. I see killer cams in the house. We got silver fox in the house. Eric, hey, it's me.
69. 12:12>> [music] >> Noemi built a an app. It's working, but she's got some work she's got to do this weekend.
70. 12:22Here's the thing about building apps and launching them in the app store. >> [clears throat] >> Then you have customers. Once you have
71. 12:26customers, now you got to deal with your customers. >> [laughter] >> So, so Noemi went from I'm no I'm no
72. 12:34technologist or I'm no programmer to I'm a programmer to holy crap, I built an app to holy crap, I launched an app now
73. 12:41she's in customer support. Laugh, laugh, laugh. >> [laughter] >> You know, you you uh you have you have these dreams of up in
74. 12:53your game and up in your game. Every time you up your game, you know, there's a step above it, right? There's a whole
75. 12:57different game you didn't know you were signing up for. Uh, wasn't it so fun to hear an AI love story today? It really was. It was so
76. 13:05good, source camp. I'm so cooked right now. >> [music] >> In a westerly direction
77. 13:20this car is my train. I've been driving, I've been wandering. >> [music]
78. 13:34>> Feel like an 80-year-old man holding on to 29. I've been here on the horizon in California land.
79. 13:54I've been feeling trucks carrying a wide load. Prefab house cut in half. I'm off to hustle my handmade goods tomorrow. Beautiful.
80. 14:05Got some economy one action going on. You got side hustle Mimi over there in economy six. Guessing economy six. Yeah, economy six
81. 14:15she's building [ __ ] She's just doing it using AI to just do [ __ ] she couldn't do before. Yeah, Sharon out there making handmade
82. 14:23goods. What's Cindy [ __ ] doing today? Cindy [ __ ] was doing some analog stuff today. Making some art in the studio. So good. So good.
83. 14:34Dr. Jay's in the house. What's happening? Maddie's in the house. >> [music] [music]
84. 14:55>> Sharon, what are you making? What are your handmade goods? Tomorrow I am so I got myself a fishing
85. 15:03license and I went and I dug out my fly rods and my boots and a bunch of [ __ ] I am going to get up at like 4:00 in the morning
86. 15:13and go fishing. >> [music] [music]
87. 15:27>> Real life stuff, no way. >> [laughter] >> Well, this is one of the things about the AI Salon. So, one of the things
88. 15:34we've learned in the AI Salon. No, one of the things we're learning in the AI Salon is that one of the byproducts
89. 15:43of having AI, which is a frictionless generation machine we as humans have always had friction. If you want to
90. 15:54learn to write a novel, first thing you need to learn to do is talk, then you need to learn to write then you need to learn story structure,
91. 16:03then you need to learn work ethics, then you need to write a bunch of [ __ ] And then you got to spend 1 to 25 years writing, right?
92. 16:14And then along comes AI and it's like I can take any idea in my head and instantly have a novel. And what we're learning in the AI Salon
93. 16:22is that there's a cost to that. That yes, it's exciting. And we as humans have not evolved fast enough yet to actually deal with that speed.
94. 16:35And so the people that are doing really interesting work in the AI Salon, one of the things they're finding is
95. 16:40this this this gift of slowing down and pulling back and doing [ __ ] in the real world. It's really cool. >> [music]
96. 16:49>> Etch glassware, laser engraved home decor barware, very cool. Very cool. Love that. Okay, so in 4 minutes, we're
97. 16:57going to be live the AI Salon or the AI Learning Lab will be in the AI Salon in 4 minutes. Slow down to speed up. That's it.
98. 17:07Just like car racing, got to slow down to speed up. Right? You got to slow down into the corner so you got control around the corner
99. 17:16so you can accelerate sooner than the other guy out the other side. That's how that works. If you get good at it, you slow down and
100. 17:24maintain your momentum. Right? So, you have slightly higher RPMs, slightly smoother turn than than the other guys.
101. 17:33Why does Max Verstappen just keep winning and winning and winning and winning? He's I don't know. He's got something.
102. 17:42He's slowing down to speed up. So, that's what we got to do in AI. Lubricant for complex systems. Yeah.
103. 17:50Yeah. Just just even time to process. What do I want? What did this thing make? What did this AI thing make? It's exciting that it made it.
104. 17:59Do I want to do anything with that? Do I want to What Gareth discovered today is that he just kept adding to his project because he could
105. 18:08and realized, "Oh [ __ ] there's never going to be an end." I never get to celebrate. He built Jarvis. He built a Jarvis system.
106. 18:21And he and he realized, it sounded like this week or last week, that he's never going to get to celebrate having done that.
107. 18:28That's [ __ ] cool. And so, he decided to just say, "I'm going to stop adding shit." And then, and then, and then.
108. 18:38So, anyway, we'll talk about that and anything you want to talk about. It's an ask me anything night. So, if you're on
109. 18:43LinkedIn or YouTube, the the link to the AI Salon is in the description. If you're on TikTok or X, it's in my
110. 18:49profile. So, find my profile, click on that, and that'll take you right over to the AI Salon, and I will see you over there in 2 minutes.
111. 18:58Uh looking forward to it, and get your questions ready. Come on over.