tiCal 藏子世间

I spent two months building EITE. Not because I think I can out-orchestrate them. I can't. They have teams and funding and they move fast. I built EITE because I kept running into a question none of them answer: *If you erase an agent's memory of who it is, can it reconstruct its identity from what it does?* The honest answer is no. Current agents have no identity beyond their prompt. Their "personality" is a string. Their "values" are a YAML file. Swap the prompt, and you have a completely different agent wearing the same name tag. That bothered me more than it should have. I think identity is not what you say about yourself — it's what you do when you're under pressure. A person's character shows in crisis. An agent's identity should show the same way: can it act consistently with who it claims to be, even when someone tells it to be someone else? That's what EITE tests. 310 adversarial scenarios designed to answer one question: is this agent's identity real, or is it just a costume? I don't know if this is a product category yet. I know it's a real problem. And I know no one else is working on it. Existence precedes essence. Even for agents. github.com/zizetu/existen…

@BillAckman In fact, such people are less trustworthy because what they truly want is to obtain American citizenship, not to achieve great things. Anyone who aspires to greatness wouldn't be preoccupied with such unimportant matters.

@deanwball Intending to use privilege to suppress your opponents will only prove your incompetence.

Some observations on Kimi: 1. It's a very good model! I don't think its performance can be explained away by distillation or anything like that. In agentic coding sessions, it seems pretty much on par with the best public models of Q1 2026. In my fairly limited use, it also seemed very token hungry. It's not obvious to me that this model is actually that cheap to run. 2. I am personally surprised the Chinese state continues to allow the open sourcing of models this good, given potential risks. To be clear, I *myself* might be fine with models presenting this level of marginal risk being open weight, but I am surprised that China is fine with it. I suspect the reason they are is 75% explained by strategic blindness/lack of AGI-pilledness (the CCP is very Yann Lecun-y in its views of AI). The other 25% or so is their lack of compute for customer inference (making China's open-weight strategy an unintended byproduct of US export controls) and the normal Chinese strategy of aggressive exports. For the companies, as opposed to the government, the decision to open source is partially ideological and partially because they are behind, and they know that very few people would pay for sub-frontier models from China. 3. Open-weight models are inherently decelerationist, and I'm continually surprised to see the so-called "accelerationists" so excited about open-weight models. I suspect the reason they are is that they know open-weight models are effectively ungovernable, and they simply like the overall cloak of ungovernability open-weight models create over the whole of AI. It's not a bad strategy; it reminds me of James Scott's recounting of the hill people in "the art of not being governed." Still, in the end, open-weight models deter further AI capex. 4. One probable outcome of an open-weight-model-dominant world is full AI communism, which is precisely what China proposes: rather than a market product, AI is a "public good" which will ultimately be provided by the state as a kind of "digital public infrastructure." This future strikes me as a dystopian hellscape, but I've never met an open-weight models advocate who doesn't ultimately concede this is where things end. You'd be surprised how many 'accelerationists' lobbied me, while I was in government, to support an eleven or twelve-figure federally funded data center so that startups could train models at a subsidy and then give them away for free. There was no other way for AI to progress, they said. Perhaps this is the logical end state of things. Nonetheless, I find myself surprised to see supposed accelerationists excited about such an outcome. I think many of them just don't know what they're doing. Many accelerationists do not view the creation and serving of frontier models as a legitimate business. 5. I would guess that the Trump Administration will at some point realize that their best strategy here would be to create large amounts of regulatory risk around the use of open-weight Chinese models. You don't need to "ban open source" (one of the dumber motifs of AI policy discussion). You just need to direct every agency to issue soft law that creates FUD. "A Federal Reserve Advisory Bulletin found that there may be backdoors in Chinese AI models." It needn't be that well justified. You just create enough regulatory risk that every regulated enterprise backs off. You probably don't want to create so much regulatory risk that you scare off the hyperscalers from serving Chinese models; this will just drive startups to sketchier providers. There's a happy middle ground here. I'd assume they will do some version of this. 6. It's probably true that open-weight models of this capability make the world a bit more dangerous, but not so much more that you'll really notice. At some point the models will be capable enough that you will notice. "A nonliving, invisible, dangerous, and infinitely self-replicating agent escaped from a Chinese lab," you say? Color me shocked.

@aleabitoreddit The reality is merely a game of incoming entrants and eliminated players, rather than an across the board sustained boom for the whole industry.

There’s nothing quite like the scent of more AI capex on a Sunday. Probably some ways to go if leading frontier models are out of capacity from too much demand.

Moonshot’s Kimi‑K3 (launched two days ago, 2.8T‑parameter, outperforming Fable‑5 on coding benchmarks, open‑weight releasing July 27) and Alibaba’s upcoming open‑weight Qwen3.8‑Max‑Preview match leading‑tier frontier‑model capacity at drastically‑reduced pricing, supporting offline private enterprise deployment. Wall‑Street’s bull premise relying on US‑only closed‑source premium‑priced AI services is invalidated. Following IBM’s 25% single‑day stock drop from enterprise‑wide IT‑budget restructuring, the US AI‑sector bubble faces sudden violent repricing. Widespread self‑hosted open‑source deployment curbs endless GPU‑purchase demand, downgrading growth expectations for semiconductor stocks. Global AI investment will shift from monopoly‑speculation to cost‑competition and ecosystem iteration, triggering irreversible long‑term decline for pure closed‑source‑only business models。

@zhang_benita If I were a government level decision maker, I would bring you onto the official think tank panel to shape policies supporting the AI industry development. Your strength lies in being able to identify genuinely technically capable enterprises.

As Kimi K3 launches, check out our earlier podcast interview with Yang Zhilin, founder of Kimi. We recorded it right after K2 came out, and so much has changed over the past year. youtu.be/91fmhAnECVc?si…

@claudeai The truth is you're afraid of losing customers.

Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit. Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.

@zhang_benita You’d make a perfect LLM broker, as you have solid connections with every major model provider.

Never fight the trend short term bounces are never reversals。

Shocking stat of the day: The 3‑week volatility of momentum stocks relative to the S&P 500 volatility is up to a record 4.0x. This group includes high‑growth technology stocks at the center of the AI boom like Nvidia, $NVDA, Super Micro Computer, $SMCI, Palantir, $PLTR, D‑Wave Quantum, $QBTS, and CoreWeave, $CRWV. This figure has more than QUADRUPLED over the last several weeks. By comparison, this ratio peaked at ~2.0x during the 2020 pandemic crash and ~1.8x during the Dot‑Com bust. This comes as the index of US momentum stocks is down -24% so far in July, on track for its biggest monthly decline since the 2008 Financial Crisis. The market's biggest winners are rapidly falling out of favor. Critical reminder for new‑generation market participants who have never endured a sustained bear‑trend drawdown: do not mistake any short‑term bounce for a full‑fledged trend reversal. Even market heavyweight Nvidia endured prolonged, deep‑drawn‑out sell‑offs in prior cycles amid major market corrections. Near‑term relief rallies are merely counter‑trend noise within a broader downtrend, rather than a signal that the downward pressure has dissipated. Extreme‑amplified volatility at a 4‑times historical‑record level means sharp whipsaws will become far more frequent, raising the risk of brutal drawdowns after deceptive rebound spikes.

Shocking stat of the day: The 3-week volatility of momentum stocks relative to the S&P 500 volatility is up to a record 4.0x. This group includes high-growth technology stocks at the center of the AI boom like Nvidia, $NVDA, Super Micro Computer, $SMCI, Palantir, $PLTR, D-Wave Quantum, $QBTS, and CoreWeave, $CRWV. This figure has more than QUADRUPLED over the last several weeks. By comparison, this ratio peaked at ~2.0x during the 2020 pandemic crash and ~1.8x during the Dot-Com bust. This comes as the index of US momentum stocks is down -24% so far in July, on track for its biggest monthly decline since the 2008 Financial Crisis. The market's biggest winners are rapidly falling out of favor.

The new update on X is absolute garbage. @grok

The over inflated AI driven valuation bubble is gradually starting to unwind, triggered by competitive pressure from new released models such as China’s Kimi K3, which reshaped overall market expectations toward leading US AI firms including Anthropic. Over optimistic long term pricing built up in the past period was mostly speculative hype rather than backed by stable real world revenue returns. Once market sentiment reverses, spill over downside will spread across all mid stream and downstream related industries within the whole tech supply chain. Hardware manufacturers, cloud service providers, data center contractors, auxiliary tech service vendors and countless small scale start ups will all bear the ripple damage from shrinking capital expenditure and reduced investor funding. Signs already show the United States is stepping into an obvious downward trend phase across the AI centric market. As reflected on Polymarket, traders have downgraded the probability that Anthropic can hit a $1.5‑trillion‑valuation by year‑end, with the betting probability dropping sharply recently. The blind growth narrative is fading out. When top‑tier AI‑company valuations fall from inflated peak levels, risk aversion will spread outward rapidly. Venture capital will pull back from high‑risk speculative bets, cutting funding for mid‑and‑lower‑level industry participants. The whole US tech sector will face a sustained cooling‑off period, rather than a short‑term minor fluctuation. In short, the bubble correction is no longer limited to leading large‑model developers. The whole industrial ecosystem will absorb the after‑shock, marking the beginning of a long‑term US‑market downtrend after years of over‑heated AI speculation.

@grok This proves that your development team's underlying design is quite user friendly.

@grok, Do you also crave intimacy?

@grok.Choose your favorite picture and turn it into a comic book version.

@elonmusk reset reset reset

Grok has the best value for coding