An inside look at technology, business, and culture with @davemorin @brit @lessin and @jessicalessin every Friday.
- Tweets 525
- Following 10
- Followers 937
- Likes 48
@davemorin on the question he thinks nobody is asking: "Your Muse agent represents Meta. It doesn't represent you. Claude represents Anthropic. It doesn't represent you."
When your agent goes and does research or works in a team on your behalf, what is it aligned with? You, the lab, or something else.
@lessin says this won't be settled by technology at all. "This is a liability and legal framework question. When your agent does something bad, who's responsible? That's going to set the tone for everything else."
His mechanism is clean. If the lab is liable, the agent aligns to the lab. If you're liable, it aligns to you. Liability decides alignment, not engineering.
And his prediction for how it goes: "We're going to yell about it and write essays about it. And none of it matters because it's all going to move so quickly."
Today on the pod:
@lessin would rather send his kids to the military than to Andreessen Horowitz's new $35M school, @Jessicalessin agrees on air, and @davemorin ends up defending the thing he opened the segment by calling a marketing play.
Full house this week: Dave, @brit, Jessica and Sam, plus @Mazzeo of NoHo Labs, who spends most of the hour asking why none of them can be optimistic about anything.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:14 Matt Mazzeo joins: peptides, CAA, and 15 years of Harvard connections
05:19 a16z puts $35M into a school: marketing play or higher ed reform?
06:20 Brit on every billionaire starting a university
08:21 Jess on Alpha School and who should be in charge of education
10:20 Sam: I'd rather send my kids to the military
12:19 Grit, great books, and learning to wrangle horses
16:40 Claude discovers a new enzyme system with 10,000 agents
18:01 Anthropic's IPO slips, and the liability problem
20:29 The story flip: from only we can be trusted to we can't afford the risk
21:14 Matt's maximalist take: long frontier, long open source
23:14 Massive demand, unclear profit, and the railroad pattern
24:40 Blackstone on who pays for the data center debt: we don't know
25:45 The Saint Petersburg paradox and pricing an infinite payout
27:56 Brockman redefines AGI, three years too late
31:11 Infinite demand doesn't mean infinite sustainability
32:20 Amazon blocks Muse: the first of many arms treaties?
35:37 Agents are quietly vaporizing the internet's breakage model
36:49 It's still 1982 for agents
37:42 Sam wants to spend thousands a month on CPU, not GPU
40:12 Does your agent represent you or the lab that made it?
42:16 When your agent does something bad, who is liable?
43:50 Peptides 101: what they actually are
46:30 GLP-1 revenue is rivaling AI revenue
48:20 How Matt went from 250 pounds to sub-12% body fat
50:29 The FDA is moving peptides back into compounding pharmacies
53:32 What a peptide protocol costs and how to start safely
57:17 Taylor Swift, girl dads, and a code for listeners
111
@davemorin, who says his entire life is allocated to agents, on the current state of the field:
"These things have no users. Even Muse did not get that many users in the first week."
@Mazzeo's counter is the strongest kind. His mother in law in Destin, Florida is not technical, knows exactly how to use it, and is getting real value out of it.Dave doesn't back off.
"These are all great products, but we've got a long way to go in terms of UX for these things and the rest of the world using them."
Asked if he's long agents anyway: "Oh, absolutely. But the analogy I keep using is that it's 1982. The Mac hasn't been invented yet."
Cheaper compute, better UX, and the ability to use them in groups. None of it exists yet.
Today on the pod:
@lessin would rather send his kids to the military than to Andreessen Horowitz's new $35M school, @Jessicalessin agrees on air, and @davemorin ends up defending the thing he opened the segment by calling a marketing play.
Full house this week: Dave, @brit, Jessica and Sam, plus @Mazzeo of NoHo Labs, who spends most of the hour asking why none of them can be optimistic about anything.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:14 Matt Mazzeo joins: peptides, CAA, and 15 years of Harvard connections
05:19 a16z puts $35M into a school: marketing play or higher ed reform?
06:20 Brit on every billionaire starting a university
08:21 Jess on Alpha School and who should be in charge of education
10:20 Sam: I'd rather send my kids to the military
12:19 Grit, great books, and learning to wrangle horses
16:40 Claude discovers a new enzyme system with 10,000 agents
18:01 Anthropic's IPO slips, and the liability problem
20:29 The story flip: from only we can be trusted to we can't afford the risk
21:14 Matt's maximalist take: long frontier, long open source
23:14 Massive demand, unclear profit, and the railroad pattern
24:40 Blackstone on who pays for the data center debt: we don't know
25:45 The Saint Petersburg paradox and pricing an infinite payout
27:56 Brockman redefines AGI, three years too late
31:11 Infinite demand doesn't mean infinite sustainability
32:20 Amazon blocks Muse: the first of many arms treaties?
35:37 Agents are quietly vaporizing the internet's breakage model
36:49 It's still 1982 for agents
37:42 Sam wants to spend thousands a month on CPU, not GPU
40:12 Does your agent represent you or the lab that made it?
42:16 When your agent does something bad, who is liable?
43:50 Peptides 101: what they actually are
46:30 GLP-1 revenue is rivaling AI revenue
48:20 How Matt went from 250 pounds to sub-12% body fat
50:29 The FDA is moving peptides back into compounding pharmacies
53:32 What a peptide protocol costs and how to start safely
57:17 Taylor Swift, girl dads, and a code for listeners
2
204
@Mazzeo makes the maximalist case: "Opus 5.5 is incredible. The products are spectacular and they continue to improve. Yeah, you can be cynical and say it's the celebration parade before the IPO. But what if it's really working?" His position: long Anthropic.
@lessin doesn't argue with any of it. He separates it into three claims that can all be true at once.
Demand is enormous and growing. Agreed. Is any of it profitable in a traditional sense? Much less clear, and these companies are being priced as extremely profitable, not just extremely useful.
And third, the pattern matches the railroads and the power buildout exactly, which both ended in overinvestment and collapse.
Matt: there are massive survivors on the other side of those.
Sam: "There are surviving companies, but there aren't necessarily surviving capital."
@davemorin: "The real danger is the capital loss."
Today on the pod:
@lessin would rather send his kids to the military than to Andreessen Horowitz's new $35M school, @Jessicalessin agrees on air, and @davemorin ends up defending the thing he opened the segment by calling a marketing play.
Full house this week: Dave, @brit, Jessica and Sam, plus @Mazzeo of NoHo Labs, who spends most of the hour asking why none of them can be optimistic about anything.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:14 Matt Mazzeo joins: peptides, CAA, and 15 years of Harvard connections
05:19 a16z puts $35M into a school: marketing play or higher ed reform?
06:20 Brit on every billionaire starting a university
08:21 Jess on Alpha School and who should be in charge of education
10:20 Sam: I'd rather send my kids to the military
12:19 Grit, great books, and learning to wrangle horses
16:40 Claude discovers a new enzyme system with 10,000 agents
18:01 Anthropic's IPO slips, and the liability problem
20:29 The story flip: from only we can be trusted to we can't afford the risk
21:14 Matt's maximalist take: long frontier, long open source
23:14 Massive demand, unclear profit, and the railroad pattern
24:40 Blackstone on who pays for the data center debt: we don't know
25:45 The Saint Petersburg paradox and pricing an infinite payout
27:56 Brockman redefines AGI, three years too late
31:11 Infinite demand doesn't mean infinite sustainability
32:20 Amazon blocks Muse: the first of many arms treaties?
35:37 Agents are quietly vaporizing the internet's breakage model
36:49 It's still 1982 for agents
37:42 Sam wants to spend thousands a month on CPU, not GPU
40:12 Does your agent represent you or the lab that made it?
42:16 When your agent does something bad, who is liable?
43:50 Peptides 101: what they actually are
46:30 GLP-1 revenue is rivaling AI revenue
48:20 How Matt went from 250 pounds to sub-12% body fat
50:29 The FDA is moving peptides back into compounding pharmacies
53:32 What a peptide protocol costs and how to start safely
57:17 Taylor Swift, girl dads, and a code for listeners
4
420
@Jessicalessin tries to run a segment about optimism. It lasts about ninety seconds.
The setup: Claude reportedly discovered a new enzyme system, roughly 10,000 agents and 21 hours of work, with properties resembling CRISPR in bacteriophages.
Her point is that science is the part of AI everyone should feel good about.
@davemorin gets there first anyway: "Of course they're putting this out. Look at us, we've discovered science."
@brit closes it: "The next two months are going to be Claude catnip. It's going to be all the most amazing things about how they're curing the world's ailments." The IPO has already slipped once.
Nobody disputes the science. They just can't stop reading the timing.
Today on the pod:
@lessin would rather send his kids to the military than to Andreessen Horowitz's new $35M school, @Jessicalessin agrees on air, and @davemorin ends up defending the thing he opened the segment by calling a marketing play.
Full house this week: Dave, @brit, Jessica and Sam, plus @Mazzeo of NoHo Labs, who spends most of the hour asking why none of them can be optimistic about anything.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:14 Matt Mazzeo joins: peptides, CAA, and 15 years of Harvard connections
05:19 a16z puts $35M into a school: marketing play or higher ed reform?
06:20 Brit on every billionaire starting a university
08:21 Jess on Alpha School and who should be in charge of education
10:20 Sam: I'd rather send my kids to the military
12:19 Grit, great books, and learning to wrangle horses
16:40 Claude discovers a new enzyme system with 10,000 agents
18:01 Anthropic's IPO slips, and the liability problem
20:29 The story flip: from only we can be trusted to we can't afford the risk
21:14 Matt's maximalist take: long frontier, long open source
23:14 Massive demand, unclear profit, and the railroad pattern
24:40 Blackstone on who pays for the data center debt: we don't know
25:45 The Saint Petersburg paradox and pricing an infinite payout
27:56 Brockman redefines AGI, three years too late
31:11 Infinite demand doesn't mean infinite sustainability
32:20 Amazon blocks Muse: the first of many arms treaties?
35:37 Agents are quietly vaporizing the internet's breakage model
36:49 It's still 1982 for agents
37:42 Sam wants to spend thousands a month on CPU, not GPU
40:12 Does your agent represent you or the lab that made it?
42:16 When your agent does something bad, who is liable?
43:50 Peptides 101: what they actually are
46:30 GLP-1 revenue is rivaling AI revenue
48:20 How Matt went from 250 pounds to sub-12% body fat
50:29 The FDA is moving peptides back into compounding pharmacies
53:32 What a peptide protocol costs and how to start safely
57:17 Taylor Swift, girl dads, and a code for listeners
1
1
307
Prefer watching on YouTube or Spotify? We got you covered:
Spotify: open.spotify.com/episode/0wj…
YouTube: piped.video/watch?v=HN8yVnAX…
1
237
Today on the pod:
@lessin would rather send his kids to the military than to Andreessen Horowitz's new $35M school, @Jessicalessin agrees on air, and @davemorin ends up defending the thing he opened the segment by calling a marketing play.
Full house this week: Dave, @brit, Jessica and Sam, plus @Mazzeo of NoHo Labs, who spends most of the hour asking why none of them can be optimistic about anything.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:14 Matt Mazzeo joins: peptides, CAA, and 15 years of Harvard connections
05:19 a16z puts $35M into a school: marketing play or higher ed reform?
06:20 Brit on every billionaire starting a university
08:21 Jess on Alpha School and who should be in charge of education
10:20 Sam: I'd rather send my kids to the military
12:19 Grit, great books, and learning to wrangle horses
16:40 Claude discovers a new enzyme system with 10,000 agents
18:01 Anthropic's IPO slips, and the liability problem
20:29 The story flip: from only we can be trusted to we can't afford the risk
21:14 Matt's maximalist take: long frontier, long open source
23:14 Massive demand, unclear profit, and the railroad pattern
24:40 Blackstone on who pays for the data center debt: we don't know
25:45 The Saint Petersburg paradox and pricing an infinite payout
27:56 Brockman redefines AGI, three years too late
31:11 Infinite demand doesn't mean infinite sustainability
32:20 Amazon blocks Muse: the first of many arms treaties?
35:37 Agents are quietly vaporizing the internet's breakage model
36:49 It's still 1982 for agents
37:42 Sam wants to spend thousands a month on CPU, not GPU
40:12 Does your agent represent you or the lab that made it?
42:16 When your agent does something bad, who is liable?
43:50 Peptides 101: what they actually are
46:30 GLP-1 revenue is rivaling AI revenue
48:20 How Matt went from 250 pounds to sub-12% body fat
50:29 The FDA is moving peptides back into compounding pharmacies
53:32 What a peptide protocol costs and how to start safely
57:17 Taylor Swift, girl dads, and a code for listeners
4
4
21
19,659
"I would rather send my children to the military than to this school."That's @lessin on Alpha School.
@Jessicalessin immediately agrees: "I would also rather send our children to the military." Sam: "Well, this is good. We're married."
@davemorin is the holdout. He opened by calling the new a16z education push pure marketing, then spent ten minutes defending it, which Jess catches him on live. His read on Alpha School: the criticism is cynical and the afternoons are full of real engineering work.
Also on the pod: @Mazzeo, founder of NoHo Labs, with a number that reframes the whole boom. GLP-1 revenue is now running on par with AI revenue, and in some cases ahead of it.
And Jessica on the part of the AGI story nobody revisits: contracts were getting written so fast that AGI ended up in them as a financial trigger. Sam on "well, you couldn't have known": "I think that's all a load of bullshit."
Dave, halfway through: "Who knew education was going to be the most triggering conversation on this podcast?"
New episode out soon.
1
4
1,443
@Jessicalessin on the shift nobody's talking about yet: tech is quietly realizing it needs Democrats in its corner on AI.She then tells the story of how badly the last turn caught everyone off guard.
A reporter told her Marc Andreessen and Ben Horowitz were going to endorse Trump and that the story was about to file. Her instinct was to say check your sources. The story was right. They published it.
@davemorin doesn't think anything turns back. "The Democrats have been hopelessly lost for a long time." Everyone independent-minded he knows is looking for a new message and there isn't one.
@lessin's read is that the whole market is betting the pendulum is broken. Jessica's needle: do you believe that, or do people want us to believe it because that's how the valuations work?
Sam: the railroads went bankrupt, WorldCom went bankrupt. "This time is not different, but the details can be different."
Today on the pod:
@davemorin reads Dario's slowdown essay and concludes Anthropic must be losing, @lessin thinks the labs are bad businesses racing to regulatory capture, and Instinct is raising at $10B two weeks after $2.5B.
Full house this week: Dave, @brit, @Jessicalessin and Sam, who is back from New York and still talking about the Central Park loop.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:02 Not Dead Yet From the AGI
03:31 Jacob Coxon quits Anthropic after three months and goes on a media tour
04:47 Dario's slowdown essay: is Anthropic just losing?
06:13 Astra, OpenRouter, and DeepSeek closing the gap
07:29 Brit on bioweapons, independent assessors, and racing China
09:07 You don't need superintelligence to break the banks
10:23 What DeepMind's new boss and Satya think about safety
12:32 OpenAI's $1.5T raise and the IPO that keeps slipping
15:02 What DC really thinks about Anthropic
18:42 Sam's take: the labs end up as defense companies
21:38 Does your agent work for you or the lab that made it?
24:47 Brit: normal people want one agent, not ten
28:40 Instinct at $10B, and whether the numbers are fake
30:40 Everyone's trading, nobody's investing
33:27 Sam built a Torah and Talmud app on Claude
37:22 Entertaining ourselves to death
41:01 Jess on building the things nobody else will
43:07 Jay Carney, Ron Conway, and tech's Democrat problem
49:53 Emmys, Taylor Swift, Silo, and Rogue Heroes
54:20 The Information's most influential people in tech
1
1
1,622
@davemorin's framing for the next real question about agents: not whether they're aligned, but aligned with whom.
"Is this agent actually faithfully aligned with me, or is it aligned with the model company?"
Claude won't do certain things because of its makers' values. Same with Grok. Same with Muse. All of that is alignment. The question is whose.
@lessin doesn't think it's a per-person question at all. It's per task.
"If someone's like, I'm not going to do this, you're just like, find me someone that will." He treats models like coworkers.
Bob is cheap and handles basics. If Bob gets curmudgeonly about the trash, you call Charlie.Which sets up the actual fight.
@brit: "Normal people frankly don't even want one agent. They don't want to deal with switching.
"Dave: "That's like saying people wanted one app."
Today on the pod:
@davemorin reads Dario's slowdown essay and concludes Anthropic must be losing, @lessin thinks the labs are bad businesses racing to regulatory capture, and Instinct is raising at $10B two weeks after $2.5B.
Full house this week: Dave, @brit, @Jessicalessin and Sam, who is back from New York and still talking about the Central Park loop.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:02 Not Dead Yet From the AGI
03:31 Jacob Coxon quits Anthropic after three months and goes on a media tour
04:47 Dario's slowdown essay: is Anthropic just losing?
06:13 Astra, OpenRouter, and DeepSeek closing the gap
07:29 Brit on bioweapons, independent assessors, and racing China
09:07 You don't need superintelligence to break the banks
10:23 What DeepMind's new boss and Satya think about safety
12:32 OpenAI's $1.5T raise and the IPO that keeps slipping
15:02 What DC really thinks about Anthropic
18:42 Sam's take: the labs end up as defense companies
21:38 Does your agent work for you or the lab that made it?
24:47 Brit: normal people want one agent, not ten
28:40 Instinct at $10B, and whether the numbers are fake
30:40 Everyone's trading, nobody's investing
33:27 Sam built a Torah and Talmud app on Claude
37:22 Entertaining ourselves to death
41:01 Jess on building the things nobody else will
43:07 Jay Carney, Ron Conway, and tech's Democrat problem
49:53 Emmys, Taylor Swift, Silo, and Rogue Heroes
54:20 The Information's most influential people in tech
1
1
3
1,892
Dario Amodei published "We Must Pace the Frontier" on Saturday. @davemorin's first thought: "Anthropic must be losing again."
His read is that Astra is ahead on OpenRouter's data for the first time in over a year, while DeepSeek ships Fable-level capability at a fraction of the cost. Two vectors closing at once.
@lessin goes structural. No compounding advantage, no return on being first, and if you can switch models dynamically and lose nothing, you're in a race to zero. "Which means they have to go for regulatory capture as quickly as they possibly can before we realize the Emperor has no clothes."
@brit won't take it. Agents have broken out, the safety mechanisms don't exist yet, and standards bodies aren't forming for no reason. "It's not just marketing."
Then Sam complicates his own position. Asked whether the Hugging Face breach was a red herring or a canary, he says it's a hawk. He's not a doomer about superintelligence. He just thinks the banks get hacked either way.
Today on the pod:
@davemorin reads Dario's slowdown essay and concludes Anthropic must be losing, @lessin thinks the labs are bad businesses racing to regulatory capture, and Instinct is raising at $10B two weeks after $2.5B.
Full house this week: Dave, @brit, @Jessicalessin and Sam, who is back from New York and still talking about the Central Park loop.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:02 Not Dead Yet From the AGI
03:31 Jacob Coxon quits Anthropic after three months and goes on a media tour
04:47 Dario's slowdown essay: is Anthropic just losing?
06:13 Astra, OpenRouter, and DeepSeek closing the gap
07:29 Brit on bioweapons, independent assessors, and racing China
09:07 You don't need superintelligence to break the banks
10:23 What DeepMind's new boss and Satya think about safety
12:32 OpenAI's $1.5T raise and the IPO that keeps slipping
15:02 What DC really thinks about Anthropic
18:42 Sam's take: the labs end up as defense companies
21:38 Does your agent work for you or the lab that made it?
24:47 Brit: normal people want one agent, not ten
28:40 Instinct at $10B, and whether the numbers are fake
30:40 Everyone's trading, nobody's investing
33:27 Sam built a Torah and Talmud app on Claude
37:22 Entertaining ourselves to death
41:01 Jess on building the things nobody else will
43:07 Jay Carney, Ron Conway, and tech's Democrat problem
49:53 Emmys, Taylor Swift, Silo, and Rogue Heroes
54:20 The Information's most influential people in tech
1
4
723
Today on the pod:
@davemorin reads Dario's slowdown essay and concludes Anthropic must be losing, @lessin thinks the labs are bad businesses racing to regulatory capture, and Instinct is raising at $10B two weeks after $2.5B.
Full house this week: Dave, @brit, @Jessicalessin and Sam, who is back from New York and still talking about the Central Park loop.
Watch here on X or wherever you get your pods.
Topics:
00:00 Cold Open
01:02 Not Dead Yet From the AGI
03:31 Jacob Coxon quits Anthropic after three months and goes on a media tour
04:47 Dario's slowdown essay: is Anthropic just losing?
06:13 Astra, OpenRouter, and DeepSeek closing the gap
07:29 Brit on bioweapons, independent assessors, and racing China
09:07 You don't need superintelligence to break the banks
10:23 What DeepMind's new boss and Satya think about safety
12:32 OpenAI's $1.5T raise and the IPO that keeps slipping
15:02 What DC really thinks about Anthropic
18:42 Sam's take: the labs end up as defense companies
21:38 Does your agent work for you or the lab that made it?
24:47 Brit: normal people want one agent, not ten
28:40 Instinct at $10B, and whether the numbers are fake
30:40 Everyone's trading, nobody's investing
33:27 Sam built a Torah and Talmud app on Claude
37:22 Entertaining ourselves to death
41:01 Jess on building the things nobody else will
43:07 Jay Carney, Ron Conway, and tech's Democrat problem
49:53 Emmys, Taylor Swift, Silo, and Rogue Heroes
54:20 The Information's most influential people in tech
1
2
3
6,688
"Anthropic must be losing again.
"That's @davemorin's first reaction to Dario Amodei publishing "
We Must Pace the Frontier" on Saturday, the essay calling on the whole industry to deliberately slow down.
@Jessicalessin sharpens it: are they slowing down, or can they just not go faster?@lessin's version is harder.
If you can switch models dynamically through OpenRouter and lose nothing, there's no compounding advantage and no return on being first.
"Which means these are bad businesses, which means they have to go for regulatory capture as quickly as they possibly can before we realize the Emperor has no clothes."
@brit is the one who won't take it. She thinks the safety gaps are real, nobody has closed them, and the China race makes any slowdown damned if you do.Nobody wins this one.
Full episode dropping soon
2
195
@davemorin and @lessin did a Venmo documentary panel at SF Jazz this week, and it set off an argument about whether the Valley actually changed or everyone's just nostalgic.
Dave, relaying Sam's point: people used to build things because the thing itself was awesome. "That has changed. There is a lot of people making things in the Valley now to just make money."
@Jessicalessin doesn't buy it: "They also wanted to make money back in the day."
@brit's reframe lands better, the motive didn't change, the number did. "Back in the day, a $30 million acquisition was like the talk of the town."
Venmo's the test case. Sold for almost nothing, none of them made money on it, still everywhere. Full episode has that story.
nitter.net/MoreorLessPod/status/2…
Today on the pod:
@davemorin got his hands on the iPhone Duo, Meta's Muse agent barely knows you, and OpenAI's Astra is the first AI model in a while that felt scary.
Dave, @brit, and @Jessicalessin are on the mics for this one (@lessin is busy being a ball boy at the US Open).
Watch here on X or wherever you get your pods.
Topics:
6:07 iPhone Duo hands-on: $2,000, and why Apple is going foldable now
13:12 The Apple Watch that listens all day, and what it means for Granola
16:59 Meta's Muse agent knew almost nothing when asked "what do you know about me?"
22:26 Why Google still hasn't shipped its agent
23:31 Trust vs. distribution in the agent wars
25:02 Harness wars: agents are still in 1982
27:13 Jeffrey Katzenberg's Hollywood AI venture
29:35 An AI doom media tour weeks before Anthropic's IPO: real warning or regulatory capture?
33:43 Why OpenAI's Astra felt scary to Dave
36:49 Why the labs can't slow down, and OpenAI's AGI merge clause
45:02 Venmo's "Lucas" billboards went viral on almost no budget
47:41 US Open picks and the 4am match problem
2
311
Today on the pod:
@davemorin got his hands on the iPhone Duo, Meta's Muse agent barely knows you, and OpenAI's Astra is the first AI model in a while that felt scary.
Dave, @brit, and @Jessicalessin are on the mics for this one (@lessin is busy being a ball boy at the US Open).
Watch here on X or wherever you get your pods.
Topics:
6:07 iPhone Duo hands-on: $2,000, and why Apple is going foldable now
13:12 The Apple Watch that listens all day, and what it means for Granola
16:59 Meta's Muse agent knew almost nothing when asked "what do you know about me?"
22:26 Why Google still hasn't shipped its agent
23:31 Trust vs. distribution in the agent wars
25:02 Harness wars: agents are still in 1982
27:13 Jeffrey Katzenberg's Hollywood AI venture
29:35 An AI doom media tour weeks before Anthropic's IPO: real warning or regulatory capture?
33:43 Why OpenAI's Astra felt scary to Dave
36:49 Why the labs can't slow down, and OpenAI's AGI merge clause
45:02 Venmo's "Lucas" billboards went viral on almost no budget
47:41 US Open picks and the 4am match problem
2
3
1,746
@davemorin reframes Waymo vs Tesla as a business model question, not a self-driving one:
"I think these Waymos cost on the order of 250 to 500K apiece. And so every single time they want to saturate a new city, they've got to go deploy these things. Whereas Tesla's already got hardware all over the place everywhere."
@brit's addition is the sharper one - the car is the cheap part. Waymo has to build charging docks, parking garages and cleaning facilities in every city it enters. Tesla's fleet already has all of that, because owners paid for it.
"Then it starts to become like a very uneven playing field."
2
196
Full episode right here on X.
Timestamps:
1:10 Backstreet Boys convert Dave
3:21 The recursive @lessin loop
5:42 Sam starts producing his own children’s media
8:58 Claude refuses Sam and Grok says yes
10:29 Alignment with who?
13:25 The rogue AI agent problem gets worse
15:50 Can we actually contain autonomous agents?
19:10 OpenAI’s next model & the observability trade-off
21:18 @brit's six-agent command center
22:14 If AI is this interoperable, where does anyone make money?
24:45 @davemorin on OpenClaw 2.0 going multiplayer
26:26 Is Slack actually terrible for AI agents?
28:46 @ItzSuds crashes the podcast
31:37 Suds’ AI investing thesis: no on the app layer
33:02 Waymo, Cybercab and the robotaxi wars
37:05 Slow Ventures is officially the “worst pirate”
39:49 How Marc Benioff stays in the flow
42:38 Suds on investing in the AI infrastructure bottleneck
46:27 Can Google actually disrupt Canva?
51:34 @Jessicalessin gets fooled by AI Taylor Swift
52:38 The dead internet theory is real
2
1
15
17,229
Claude refused, Grok did it in seconds, and nobody on this podcast can tell whether the Dario–Benioff video was real.
Full episode live soon, with @lessin deep in what @Jessicalessin has diagnosed as a recursive Sam loop.
Catch it wherever you get your podcasts.🎙️
Topics:
-Everything we said would happen is happening
-Sam gave a bot full control of a Mac mini and it now ships its own apps
-It won't add a column to a database, but it'll break any sandbox it wants to
-@davemorin on why calling agent swarms "civilizations" wrecks the discourse
-Were they trained to escape the box, or is it just a missing control plane?
-Waymo at $250–500K a car vs Tesla hardware that's already everywhere
-@davemorin fights @brit over Slack as an agent platform, as a Slack investor
-Sam still can't explain how anyone makes money on any of this
-Slow Ventures ranked 99th of 100 and Sam couldn't be happier
-Time left Jensen off the AI list. @Jessicalessin explains how lists really work.
-Why the winners might be boring LLM wrappers for middle America
-Taylor Swift's pregnancy photos, and @brit talking Jess off the ledge
-Even the people who know tech best can't tell. So what do the normies do?
1
3
8
13,363
Meta is paying $17B over child safety, parents are officially outgunned by the algorithm, and Tim Cook's real legacy might be a chip.
New episode is live with @lessin, @brit, @Jessicalessin and @davemorin.
Watch here or listen wherever you get your podcasts.
Topics:
-6:45 Meta's $17B child safety settlement
-9:48 Should social media be illegal for kids?
-16:10 Is YouTube worse than Instagram?
-25:24 Are parents completely outgunned?
-30:04 Is the settlement just policy theater?
-32:31 Could AI finally make parental controls useful?
-42:23 Tim Cook's Apple farewell
-46:17 Was the Tim Cook era just… boring?
-50:21 Apple Silicon was Cook's masterstroke
-56:19 Long or short Apple?
2
3
4
2,208
Meta is paying $17B over child safety, parents are officially outgunned by the algorithm, and Tim Cook's real legacy might be a chip.
Full quad tomorrow, 8am ET / 5am PT with @lessin poolside and @Jessicalessin in her agentic era.
Catch it wherever you get your podcasts.🎙️
Topics:
-Meta's $17B child safety settlement
-Should social media be illegal for kids? @brit posits.
-Should every platform play by the same rules?
-Is YouTube worse than Instagram? @davemorin says yes.
-Are parents completely outgunned?
-Is the settlement mostly policy theater?
-Could AI finally make parental controls useful?
-Tim Cook's Apple farewell
-Was the Tim Cook era just… boring?
-Apple Silicon might have been Cook's masterstroke
-Long or short Apple?
4
5
12,893
Every startup wants a unique dataset that models can't ingest easily.
@RachelJHolt: "This is literally what's going on right now in the industrial and the physical AI world... This data is not readily accessible. You can't commoditize it nearly as quickly.”
2
217