Not is it only dumb, but it is plain unimplementable. Are they saying the HMI interfaces on CNC machines need to be able to parse the GCode generated by any of dozens of CAM software options out there and divine if it might be gun related? That is not possible.

Theo can sleep tonight.

I don't think so. There are other knobs they can tweak to reduce load that affect quality less than quantizing. Like trimming the conversation length without telling you, reducing reasoning effort, etc.

We never do anything that reduce model intelligence like that

You said "like that", ok but there may be some truth to reduced model intelligence. Also how AWS deployed Anthropic models for Amazons Kiro feel much dumber than those controlled entirely by Anthropic. Can't be just me

The One China policy is a fiction of foreign policy statecraft, designed to sideline the issue without having to actually deal with it. It is quite clear that apart from the official fiction there is a real policy that is not One China. This is made clear by the weapons sales to Taiwan that specifically calibrated to make a Chinese military action harder.

If it's a fiction https://www.congress.gov/crs-product/IF12503 then it's a very deceitful and and elaborate one.

No, they are not dead. However, they face incredible competition in a brutally commoditized product space.

AFAIK in some space they're still the best models on offer.

The way I see it, this was the case until a few months ago. Today, Opus 4.5 is just as good or better than 5.2 Pro at tackling hard questions and coding, Gemini beats the free models, and Kimi K2/K2.5 is the better writer/editor.

In my own testing these models sill have a different flavor to them

- Opus 4.5 for software development. Works faster, and tends to write cleaner code.

- GPT 5.2 xHigh for mathematical analysis, and analysis in general (e.g. code review, planning, double checks), it's very meticulous.

- Gemini 3.0 Pro for image understanding, though this one I haven't played around with much.

Not in my experience, Gemini proves much better for me now.

Can you get Gemini to stop outputting code comments yet? Every single time I've tried it, I've been unable to get it to stop adding comments everywhere, even when explicitly prompting against it, seems like it's almost hardcoded in the model that code comments have to be added next to any code it writes.

This is why: https://arstechnica.com/gadgets/2026/01/core-ultra-series-3-...

Intel doesn't have any spare capacity.

Is it just me, or has Claude Code gotten really stupid the last several days. I've been using it almost since it was publicly released, and the last several days it feels like it reverted back 6 months. I was almost ready to start yolo-ing everything, and now it's doing weird hallucinations again and forgetting how to edit files. It used to go into plan mode automatically, now it won't unless I make it.

Exactly. There is a big difference in code quality with state-of-the-art models versus 6 months ago. I'm strongly resisting the urge to run Claude Code in dangerous mode, but it's getting so good I may eventually cave.

The difference here is the qualitative difference that has existed between Google Search results and other competitors. Switching away from Google Search is a high friction move for most people. I'm not sure the same goes for AI chat.

I already don't use ChatGPT. I use OpenWeb UI with OpenRouter, and the API costs for my usage are peanuts. Switching to a different interface is so easy many people will. (You don't need to self host. T3 Chat, for example.) This is the difference between Google Search and ChatGPT.