Where the AI market is talking

Review: Chuwi's $449 Unibook laptop is a funhouse-mirror MacBook Neo

>The Unibook’s keyboard is pretty good, and its trackpad is pretty bad. Just like 98% of non-Apple laptops. Amazing how after all these years nobody could replicate the feel of macbook's trackpads.

Our decision on Cursor following its acquisition by SpaceX

> As AI capabilities advance, we also have a new level of accountability to ensure our upcoming model, Astra, is being used in accordance with our terms. I think Sam Altman has played a clever business move here, they can use this excuse immediately after the Hugging Face leaks and it's not going to be dismissed willy-nilly. Whereas long term they probably wanted to end their partnership with Cursor anyway. I bet Dario will be happy about this too, but maybe the compute deal prevents him from doing similar anytime soon. Very interesting times we live in...

Bootstrappable Builds: How and Why

I was thinking an FPGA with an open toolchain might make a pretty solid hardware seed that can fairly easily be replicated, without needing your own fab or anything completely impractical. In theory the FPGA could still try to recognize and backdoor a bitstream right when you load it, but practically there's a whole bunch of different open-source cores that you could load on the FPGA, and it would be very hard to try to recognize and backdoor all those potential designs at the netlist level, or at least it would take a suspiciously large area, potentially visible as clearly more complicated than what should be needed on a microscope. You could maybe distribute these verified boards that are just the FPGA and the verified CPU core + kernel + software seeds all stored on a basic NAND flash.…

The Hugging Face attack surprised me

Nvidia agrees to buy Hugging Face for $12.9B, report says

GUIs should be fully keyboard-driven

I'm a keyboard based UI lover, but I wonder if the supportive crowd is, and will ever, be too small. Even crude AS400 days keyboard UIs were beyond nice. But then if people want to try and replicate emacs keymaps or vi command composition, better for the dozens of us. ^^

Run Qwen3.8 27B locally: real numbers from my Mac Studio

This is a problem that Hugging Face could easily help solve. Crawling through Reddit or forums to find the right incantation to run a model is frustrating.

Nvidia turns Hugging Face checkpoints into C++ inference

Run Qwen3.8 27B locally: real numbers from my Mac Studio

That isn’t right. Use oMLX or something similar to serve the model with internal MTP enabled. You should get at least 40 tok/sec, but I’m not sure what hardware you are using. Dflash will mess up batching, not really worth it. You need to sift through hugging face for the right model though, and some of the MTP models are meant for rapid MLX and not oMLX.

Humanity has the debate about AI consciousness backwards

It can REPLICATE better thought out arguments it learned from the internet, which were thought by people.

Ask HN: Hugging Face is out. Who is hosting open models?

Ollama, Replicate, Together. Ollama is probably the most advanced but smallest library. I think Ollama will probably step in and replace HF because they have always been first movers. They did llama.cpp better than llama.cpp did llama.cpp and were very quick to bring highly usable prosumer solutions to market. Why not GitHub or one of these wannabes like Tangled? Seems perfect for Tangled - why not?

Decompiling a Nintendo 64 game in 84 days

Matching decompilation is a verifiable target that ensures feature parity but it says nothing about being reasonable non slop decompilation. Disassembler output can relatively easily match original binary, I doubt anyone would consider it decompilation. Let's say you add requirement of being higher abstraction language than assembly. You can translate the assembly code 1:1 with simple string substitions to C code which would provably replicate the behavior of program perfectly. I would consider that more of hard coded emulation, opposite of JIT like ahead of time code translation. Just because the intermediate langauge is higher level language than assembly doesn't automatically mean that the program text operates at higher level of abstraction than assembly. I have seen plenty of decompi…

Trade (and Tariffs)

1) (all) tariffs aren't obviously and universally bad, only when applied universally without any coherent long term goal/strategy. 2) retaliatory tariffs are targeting the hostile actor, not widely against all (other) trading partners, thus redirecting trade towards less hostile country.

Humanity has the debate about AI consciousness backwards

I am telling you to ask a chatbot because you are saying things so hilariously philosophically incompetent and ignorant that even a retarded LLM will improve your thinking. You obviously don't care about thinking coherently though, or can't even recognize what coherent thought looks like (mainly because you seem to think something along the lines of "my few minutes of armchair reflections prove all philosophy is vague and retarded"), so this is clearly a waste of time. I will not be engaging further.

Small Models Have Arrived

It does have to be said that if LLMs keep becoming better coders at some point the bottleneck on quality is prompting. Good ideas have many hidden assumptions you think are procedural but often are pivotal to your broader vision. I find that when I give an LLM my full handcrafted codebase, it does very well. It follows my conventions, sees the intent and can coherently build within its scope. It writes much better code than a 'vibe' prompt. It is always tempting and I myself will continue pushing the boundaries, but when you keep an LLM in reasonable scope (that may be one line, function, file at a time, depending on your idea of reasonable), you, by definition, can get sound utility out of them.

Canada will match US tariffs 'dollar for dollar' as trade talks break down

Except I haven't devolved from anything. It's entirely consistent, and the whole chain confirms that it isn't possible to coherently debate this with people whose starting position is insane, like yours. At it is insane to assert that "trivially every other country has a better rack record" than the US does, and it is insane to claim that the US was founded with the express intention of killing off the indigenous inhabitants of North America. These are insane positions to hold, and yet ... here you are.

Trade (and Tariffs)

It isn't just about that. You need industrial policy. And not just stuff like the CHIPS act. In China a small outfit can get financing and open a factory anywhere. You can't do that in the USA. You are too small for the lenders or investors who do such things to care yet too risky for traditional financial institutions to risk anything on you. If I want to buy a couple of CNC machines and start making a product in my garage I have to hope I have enough room on credit cards to do it because I cannot get the money from anyone. Doesn't matter if I have a good idea. Doesn't even matter if I have some initial customers/orders. If the government is going to piss away money then at least piss it away on that sort of economic activity. If you can present a coherent business plan or product d…

Silicon Valley is in denial in face of widespread backlash

I am curious how that will be coherent but we have all been surprised by incoherence in politics many times over.

Anthropic Is Quietly Cutting Google Out of the Equation

Nvidia projects $673B in sales as AI demand widens

Counterpoints: 1. Depending on the data source you look at, about 50 - 60% of people use AI at work but only for 5 - 15% of work hours. That leaves about 2x (from users) times 7 - 20x (from work hours) for growth. Furthermore agentic usage is much more token-intensive than regular prompts, that's another unknown multiple that will get applied. Small models will make a dent for sure, but even they need to run on hardware. It's not clear how much their lower resource requirements will cancel out the scope for growth, but I think it will take time for that dynamic to play out; people are only just starting to ease up on tokenmaxxing. Anthropic revenues would be the canary in the coalmine, and thankfully they'll be IPO'ing soon. 2. All the relevant fabs (mainly, TSMC) are extremely capacity-c…