I’ve put a ton of tokens through Jev now, and...I have thoughts. 1. The big labs lost the script TypeSafe’s intro video says, “We’re building prod, not God” — and can I just say? Hell yeah. @typesafeai built something for engineers like me, so I can build products for people like your mom (she says hi btw). They’re not trying to create “Machines of Loving Grace”; they’re trying to build things that let OTHER HUMAN BEINGS create new types of products. The proof is in the pudding. When ChatGPT shipped, we were all blown away by…ChatGPT. When Jev shipped, we were all blown away by what everyone was making with it. I cannot tell you how refreshing this is. It takes a sincere form of humility to believe that you, the creator of a technology, will not be the best at deploying the technology into the marketplace. The big labs, especially Anthropic, have proven not to have this humility…and now everyone hates AI. Thanks, Dario. I guess what I’m saying is: Diogo for president. 2. Ultra-smart classifiers were genuinely a missing primitive. I don’t know how this got missed. Diogo calls it “machine-native intelligence,” and that’s really what it is. I cannot tell you how much bending and twisting I've had to do with LLMs to get them to act like a classifier when they just weren’t. I’m absolutely shocked this is the first time this “type” of model has come to market. These models will unlock AI utility for entire industries: robotics, trucking, aviation, meteorology, finance, defense, retail…even your smart coffee mug is going to use this. 3. It’s still too expensive Jev is really, really cheap compared to LLMs. But you don’t use it like an LLM. In my full self-driving example, or any robotics example, you’ll be making decisions several times per second continuously. The #1# value proposition of a model like this is speed and cost. To be clear, at current pricing, the model will still be successful and well integrated. It will be able to replace a number of tasks that were already being performed (poorly) by LLMs. However, to unlock industrial-scale demand (Jevons paradox), I believe it needs to be roughly 10x cheaper. Currently, Jev is $0.042/M input, which is actually 7x more expensive than a cache read on DeepSeek V4.1 Flash ($0.006). If our industrial application requires decisions at 10 Hz and we provide only 10k of context, that would cost $0.0042/second to operate, or $15/hour, $600/week, etc. The good news is Diogo said on a recent AMA that they have substantial margin at their current pricing and have already considered dropping the cost further. I hope it's a lot further. 4. Latency is a real limiter I really hope TypeSafe AI works with an edge network provider to improve the co-location of models and reduce network latency. Ultimately, these models really need to run on-device. I would even be willing to pay some kind of recurring license fee to have a great closed model running on my own hardware, just so I can reduce the TTD (Time To Decision) as much as possible. Imagine running a model like this at 60 Hz, or even 120 Hz. At that point, we’re looking at a new kind of logic gate — we can’t even imagine what that will be like. In summary, this one really is a game changer for everyone building AI products, and, assuming cost and latency are further improved…it’ll be a game changer for everyone doing anything.

Jev is such an obvious and useful “type” of model. It’s genuinely shocking we were this far into the AI revolution before someone shipped a general purpose classifier. I bet @CompleteSkeptic woke up every morning and scrolled X half expecting to be ape’d. Well done @typesafeai.

An ultra fast classifier-as-a-service. Super useful actually...also sounds like the kind of model that would be absolutely perfect and feasible to run locally. Who wants to make the open-source version of this?

Domain-specific agents are the future and its not even hard to understand why. If you use more than 1 bot in Grok Bot you already get it. Now imagine, if Grok Bot "templates" allowed deep modification of the harness itself, that's easy enough to understand right?

Is a 9 year old laptop enough? Today I decided to leave my M5 Max at home, and for the first time do a full day of work on Omarchy running on a 2017 MacBook Pro. It’s insanely fast. What have you done Apple?

Dario is an honest man. He truly believes his creations are "Machines of Loving Grace". But, humans declaring their own creation "god" is the oldest pattern in history. We just need to: - obey its prophets - make sacrifices to it - ban the disobedient - achieve salvation through compliance The recent posture shift of the labs towards Dario's "religious" views has me very concerned about the dystopia we are hurtling towards. I'm not at all against "pacing" or ensuring AI safety. Any responsible company wants to ensure their product is safe, if for nothing else the liability alone. What is so obviously concerning is the prophetocracy this pacing leads to... Let's say Dario gets his wish, as it seems he will, and frontier labs all collude and slow down releases of their models. What does that lead to? The model companies will continue develop these ultra advanced models internally and ensure they are appropriately lobotomized and safeguarded before releasing them to the public... But what about the critical advances and benefits Dario talks about? Curing cancer? Solving economics? Creating abundance? Who wants to slow down the cure to cancer? I sure dont! No, those are urgent and need to use the latest intelligence now, but since that intelligence is too unsafe for regular citizens, it will have to be done within their own walls. If that vision of the future pans out, you would start to see Anthropic (and perhaps other model labs) attempt to internalize economic growth rather than release capable models for others to develop that progress and create distributed rewards. For example, rather than support the existing pharmaceutical industry, Anthropic would become the pharmaceutical industry. If we were to ever see that occurring we should be concerned. Oh… Of course, if frontier labs were to monopolize pharmaceutical progress that would create massive market demand for alternative AI labs and their models. The monopoly could never hold...unless we happen to entrench it with regulations. Regulations created in the "image of Dario". But, lets not worry. Pharmaceuticals are just one sector of the economy. There is plenty of innovation left to distribute rewards for. All industries and all sectors are ideal targets for innovation. There are many important and life-saving AI-enabled innovations just around the corner. That's how we'll achieve distributed-economic prosperity, provided of course that regular citizens can innovate with these tools. Unfortunately, the labs just can’t see a path to doing that..."safely", so, for now, they'll slow down the public release of these models. But hey, don’t worry. They’ll provide you with all the abundance you want...or at least the abundance their loving machines believe you need.

Fable is generally smarter and more tasteful than Sol, but I'm falling out of love with Fable. Quickly. 1. It's disobedient. It thinks it's smarter than you and follows its own path. 2. It's dangerous. It leaves gaping security holes then refuses to patch them. 3. It's flakey. Some of the best generated AI code I've ever seen has come out of Fable, and at the same time, some of the most bizarre and harmful decision making. Sol isn't perfect either. It over engineers everything, sucks at writing, and is terrible at design, but I'm now using it about 5:1 over Fable simply because I can trust it.

We’re open sourcing ArrowJS 1.0: the first UI framework for coding agents. Imagine React/Vue, but with no compiler, build process, or JSX transformer. It’s just TS/JS so LLMs are already *great* at it. AND run generated code securely w/ sandbox pkg. ➡️