Amjad (Replit CEO) cryptically responds to the Fable export control situation: 'Feels like we're getting psyoped. The end-game here is something bigger.' 2292 likes—the highest engagement tweet of the day.

The suggestion is that the Fable restrictions aren't really about cyber exploitation risk—they're a precedent-setting move for broader government control over frontier AI models.

Levie's comprehensive analysis of the Fable export control situation as net positive for regulation discourse. He frames it as 'an early peek into what AI regulation would end up looking like at scale when enacted at the model layer instead of the specific application.'

The core problem: the government would have sole discretion over when a model can be released, based on factors they control. You'd debate risks for months with every release. 'If this paradigm had existed 3 years ago at the start of the current AI wave, we'd likely currently be stuck on GPT-4 level intelligence.' He advocates for regulating applied use of AI—not models. 318 likes.

Levie on why the model routing layer will increase in value substantially. Three reasons: (1) Cost optimization—frontier for planning/review, cheaper model for bulk work. (2) Capability maximization—different models excel at tool use, coding, knowledge work. (3) Risk mitigation—the Fable situation proves you need flexibility to deploy across providers. 'It's going to increasingly be a strategic advantage for the applied AI layer that they can effectively route between models.' 209 likes.

Garry Tan's call to action: 'In AI most people are still trying to use old maps on a new territory. Throw the maps away. It's time to draw new ones. The only way you can do it is walking the land.' 1154 likes.

This is the meta-theme of the current moment: every framework, every best practice, every mental model from the pre-AI era needs to be questioned from first principles.

Steipete's viral Codex story: 'Got a PayPal verification text and thought I been hacked, but it was just Codex signing up for a web service it needed.' 1371 likes.

This is the autonomous agent era in a nutshell: the model is now taking actions in the real world—registering accounts, making verifications—that were previously exclusive to humans. The line between 'tool' and 'agent' has fully blurred.

Thibault Sottiaux (OpenAI) jokingly introduces himself to the Codex community: 'Hi, I'm Tibo and I just discovered Codex. AMA.' 1442 likes—the second highest engagement of the day.

A moment of levity from the OpenAI Codex team lead during an intense week of Fable regulation news and competitive posturing.

RealMadhuGuru's empathetic inside perspective on frontier model launches: 'Launching an LLM isn't like shipping traditional software—you're making a decision about a black box with effectively infinite use cases and infinite failure modes.' Every increase in capability expands both valuable uses and potential misuse. You build evals, red-team, iterate on candidate checkpoints—then early-access partners still uncover behaviors you didn't anticipate. 'You can never be 100% certain you've understood a frontier model.' 29 likes.

Dan Shipper shares his before/after Fable ban usage: Claude app vs Codex app usage shifted dramatically after Fable was suspended for foreign persons. 146 likes.

A real-time data point on how regulation immediately redirects user behavior—the ban didn't reduce usage, it just moved it to a competitor.

Nikunj's VC transparency: 'Not sure which founder needs to hear this, but paid partnerships and boosting views on X are now negative signal in all VC group chats. I'm hoping this comes to the inflating ARR trend soon, but that might still take another cycle.' 98 likes.

Steipete flags: 'This shortage of chips is getting out of hand.' 512 likes.

The chip shortage is becoming the physical bottleneck for the entire AI industry—and it's hitting individual builders, not just big companies.