AI's new bottleneck: taste. eBay's corpse.

AI's new bottleneck: taste. eBay's corpse.

The Throughline Brief · June 28th. 5 min 35 sec.

Cold open

An internal OpenAI model just disproved the complex Erdos unit distance conjecture. It did this by thinking. For a long time. The old benchmarks are completely broken. Today, an AI capability is directly tied to how much money and time you let it spend chewing on a problem. The frontier of intelligence just shifted.

Intro

This is The Throughline Brief for June 28th. I went through every big AI and tech podcast this week so you do not have to. Here is what actually mattered. I am West. Let us get right into the signal.

Taste Is The Bottleneck

AI is completely flipping the economics of building products. We are seeing a massive cost inversion across the board. Implementation is becoming cheap. Taste is becoming the new bottleneck. Paul Erlang from FOMO went on 20VC this week to lay out the math. He and his team built a complex trading app in a single month. That pace is purely from AI workflows like Cloud Code and Codeax. These tools drive massive acceleration to engineering output. A small team can now do the work of a massive org in just a few weeks. Think about that. Here is what that means for you as a founder or an operator. When the code itself is cheap, the hard part changes entirely. Andrew Ambrosino leads the Codex product at OpenAI. He told Lenny on Lenny's Podcast that the actual building of software has become the absolute cheapest part of the equation. You can generate a multitude of prototypes in a single afternoon. The scarce skill is curation. You need the taste to sift through all that AI generated output, identify what is actually good, and refine it into something users actually want to buy. The human in the loop is shifting from creator to editor. If you are building a product right now, you need to hire for taste. You need people who know what excellence looks like. The AI will give you average a hundred times over. Your use is in the curation.

Vision Over Inertia

Pushing a massive vision through corporate inertia is brutal. The Acquired podcast did a deep dive on Walt Disney this week to prove it. When Disney dreamed up Disneyland, his own board of directors hated the idea. They looked at it and saw a massive financial risk. They saw a silly amusement park. So Walt bypassed them completely. He funneled his personal resources into a separate company just to fund the initial planning. He licensed his own intellectual property to himself to get it off the ground. He had to build a shadow company to force the vision into reality. We saw the modern version of this on the All In Podcast. Ryan Cohen broke down his 56 billion dollar bid for eBay. Cohen looks at eBay and sees a stagnant giant. He sees professional management trapped in corporate inertia. He wants to tear that down. He built Chewy into a multi billion dollar business through fiercely aggressive efficiency. Now he wants to apply that same sledgehammer to eBay. He is daring them to adapt to new market realities. As an investor or operator, the takeaway is stark. If you want to build something generational, expect the board to hate it. Expect the professional management class to push back. You have to force the issue. You have to back your own conviction with real capital.

Dr. Layne Norton

Dr. Layne Norton went on the Huberman Lab podcast to talk about supplementation and fitness. They got deep into the science of muscle growth and what actually moves the needle for human performance. Norton was incredibly blunt about the supplement industry and its marketing hype. He said if you are using any other form of creatine I think you are wasting your money. That is a massive statement. He is directly challenging millions of dollars in supplement marketing that tries to sell you complex, expensive variations. For anyone tracking health optimization, the science points straight back to basic creatine monohydrate. Other variants are simply a drain on your wallet.

Ryan Cohen

Ryan Cohen sat down with the guys on the All In Podcast to talk about his massive plays in the market. The conversation turned to his staggering 56 billion dollar bid to take over eBay. He explained his intense drive to acquire such a massive, stagnant target and force an absolute transformation across the entire company. He told the room, life is too short to do it small. That captures his entire thesis as a founder and an investor. You build generational wealth and shift markets by taking massive swings at the entrenched giants. You have to force the issue.

Mark Chen

OpenAI Chief Research Officer Mark Chen was on Latent Space this week discussing the future of model architectures. They were exploring how human researchers are going to transition into high level vibe researchers as autonomous models take over the heavy lifting of innovation. During this breakdown, Chen casually dropped an incredibly weighty prediction. He asked the hosts, AGI is coming soon, right. Hearing the top research executive at OpenAI say that out loud changes the timeline for every single builder listening. He is signaling that artificial general intelligence is imminent. Every roadmap and every product strategy needs to account for that reality right now.

The bottom line

Artificial intelligence is completely rewriting the rules of product development and model evaluation. It is bringing incredible use to small teams who possess elite taste, forcing a massive shift in how value is created. Meanwhile, ambitious founders are actively challenging corporate inertia to build the next era of tech giants. You have to move fast, curate ruthlessly, and take massive swings. That's your Throughline. See you next Sunday.