We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
The most actionable pick is Aaron Levie’s pointer to Jeffrey Huber’s formulation that AI is “better at brute forcing than you”: Levie treats it as a way to find agent opportunities where tireless, exhaustive work was previously impractical.
The standout: a practical test for agent opportunities
- Title: “AI is better at brute forcing than you” (the opening line of the post; no formal title is supplied).
- Content type: X post.
- Author/creator: Jeffrey Huber.
- URL:Original post.
- Recommended by: Aaron Levie, who calls it a “pretty good way” to think about the value of agents and where the opportunity lies.
- Key takeaway: Agents can tirelessly perform work that was valuable but not remotely practical before—such as simulating every security risk, reading every contract, or combing through a customer base for upsell signals. Levie’s startup heuristic is to target markets where adding compute qualitatively changes what customers can do.
- Why it matters: This is a usable screen for agent ideas: look for domains where exhaustive analysis is valuable, repeatable, and previously too costly or laborious for people to carry out.
Two design-side correctives
Stop the AI Confidence Theater
- Content type: Post.
- Author/creator: Elena Vera.
- URL: Not supplied in the recommendation; context in the Lenny’s Podcast interview.
- Recommended by: The Lenny’s Podcast host, who describes it as a “really great post.”
- Key takeaway: The post is presented as a critique of people broadcasting AI accomplishments as if they have figured everything out, even though few of those examples are actually working well. The surrounding advice is to be humble, keep experimenting, and stay curious as the tools change.
- Why it matters: It is a useful guardrail against mistaking polished AI success stories for a repeatable operating method.
The Design of Everyday Things
- Content type: Book.
- Author/creator: Not named in the interview passage.
- URL: Not supplied in the recommendation; context in the Lenny’s Podcast interview.
- Recommended by: Ian Silber, head of product design at OpenAI.
- Key takeaway: Silber calls it a classic that shapes how he thinks about design and says it is worth reading whether or not someone works in tech.
- Why it matters: Among AI-specific recommendations, this is a reminder to return to durable design principles rather than treating new tools as a substitute for understanding how products should work.
A contextual governance recommendation: The Vision of the Anointed
- Content type: Book.
- Author/creator: Thomas Sowell.
- URL: No direct book link is supplied; recommendation context in David Sacks’s post.
- Recommended by: David Sacks, who invokes the book while arguing about the concentration of authority over frontier AI.
- Key takeaway: Sacks says the book captures the mindset of “elite intellectuals” convinced that only they are enlightened enough to control the outcome, and uses that framing to criticize concentrating power in a small group.
- Why it matters: It offers a political-philosophy lens for the central governance dispute Sacks identifies: whether frontier AI is too powerful to distribute or too powerful to centralize.
Jason Lemkin (@jasonlk) recommended a video of Klaviyo CEO Abi Abialecki's SaaStr AI talk on how a 2,300-person, $1.5B ARR public company builds with agents — the one he'd most want a founder with an aging codebase to hear — with a YouTube link. The takeaways he called out:
- Every employee must hit AI skill 'L3' (running a team of agents); everyone from the CEO to summer interns commits code .
- Klaviyo builds agents with agent teams: the 'Dark Factory' prompt acts as PM, decomposes problems into subsystems, and writes contractual API interfaces; Composer hit 95,000+ users in month one after a one-weekend run .
- The base model is a 'very athletic high schooler'; the harness — a proprietary data feed plus a coach agent — scores every campaign before it ships .
- Customer agents ship pre-trained on 5-10 of their own use cases at a 50-70% resolution rate .
- Headless is the default: Twilio was 'left for dead' 18 months ago growing 4%, now almost 20%, because its API works with agents . A follow-up post quoted Abialecki: 100% of Klaviyo PMs are L3s and all commit code .
Resource recommendations from Lenny's Podcast:
- The host recommended Elena Vera's post "Stop the AI Confidence Theater" as a great post that criticizes people posting AI accomplishments and pretending they have it all figured out, noting few of those things are actually that great and working .
- Ian Silber, head of product design at OpenAI, recommended "The Design of Everyday Things" as one of the books he recommends most to others, calling it a classic of how he thinks about design and saying "You have to read it if you're in tech or not in tech" . He added that it's not a unique recommendation and that few purely design books have stuck with him .
DoorDash co-founder @stanleytang endorses a video by @andyfang (Andy Fang) on agent-first commerce: an AI agent watches your pantry shelf and reorders the moment it runs low, with builders now using the DoorDash CLI. He calls it 'insane and cool' to see in real life. The tweet embeds the video (https://video.twimg.com/amplify_video/2080769801157480448/vid/avc1/2560x1440/Yl3kMpXfCadpwOsN.mp4?tag=29).
David Sacks (Craft Ventures, All-In co-host) recommended Thomas Sowell's book "The Vision of the Anointed," citing it as capturing the mindset of "elite intellectuals convinced that only they are enlightened enough to control the outcome." He invoked the book in his critique of Anthropic CEO Dario Amodei's AI-regulation arguments, using it to argue that concentrating power in "an enlightened few" has rarely produced the promised results .
Garry Tan (@garrytan) recommended an Instagram creator's reel, calling him "amazing" and urging followers to follow him, with the reel at https://www.instagram.com/reels/Db9H43evBi4/.
Aaron Levie (@levie) recommended an X post by @jeffreyhuber as "Pretty good way to think about the value of AI agents and where the opportunity lies" . The post states: "many of the gains of AI boil down to: AI is better at brute forcing than you" . Levie elaborates that AI agents can tirelessly do things previously impractical — e.g., simulating every security risk in a codebase, reading every contract, or combing the customer base for upsell data — and the real opportunity is to "throw intelligence at [problems] that won't give up," targeting markets where more compute qualitatively changes what customers can do . Link: https://x.com/jeffreyhuber/status/2089020697876275202.