bashtoni

My experience running quantised models is that they look the same at first but are all noticeably worse once the context size gets beyond anything trivial.

No, this is not a small model. I believe 4x Sparks is possible but 4.1 is pretty big for a flash model.

MiMo 2.6 Flash is the most capable model you can currently run on 2x Sparks without resorting to a quantised version of something.

"a minor update" they say before showing off benchmarks that display a substantial improvement.

You might not buy any of his products, but by visiting X you are a product he's selling to advertisers. The fact you block ads is not disclosed to the advertisers he sells to, but your regular visits absolutely are.

Russian social media app has backdoor. Who would have thought?

(Yes, I know they're technically Dubai based now)

Is AI going to put Mathematicians out of a job, or is it going to create many new jobs reviewing proofs it creates?

I'm not sure it's clear right now.

I love how much inside information Xiaomi are giving us on the every day challenges faced by AI labs. Before the release of MiMo 2.6 it was the RL dashboard, now this update.

Had to chuckle at this when outside the US the problem with EVs is that there are so many options people find it tough to choose.

Denza or Zeekr or Genesis or Tesla? Xpeng or Leapmotor or Geely or Kia or Hyundai or Mazda? BYD or MG or GWM or Chery or Jaecoo?

Plus the various Porsche, Audi and VW electric options, Subaru and Toyota, even the new Range Rover. There are far more electric options than petrol, and a lot of them are significantly better at the same or lower price point.

NZ isn't big enough to think about. Don't tell them because they get very upset.

Don't be evil -> What does "evil" really mean, anyway? -> Judgements about good and evil don't belong in business -> Evil is pretty profitable, actually

Are the Artificial Analysis benchmarks really worthwhile any more?

They really don't seem to match my real-world experience, and based on the comments I see I don't think that match most other people's either.

For example, Opus 5 was at the top for some time. My experience is that it's not noticeably better than Opus 4.8, and it definitely seems worse than Fable 5, which AA benchmarks put behind Opus 5. GPT 5.6-sol and Opus 5 seem pretty interchangeable, although Sol is noticeably better at finding problems in code, particularly edge cases.

Based on my observations, the most important characteristics of a CEO is to be over 6ft, and to walk about looking important.

CEO jobs are safe until the humanoid robots have mastered this skill.

COO just resigned too. Coincidence?

https://xcancel.com/bradlightcap/status/2087211567012032862

The short version I give to non-technical people who ask me about whether "AI will replace coding" is this: it accelerates you. You can get much further much more quickly.

If you don't know where you're going or how to get there, or even if you're just not paying enough attention, it will get you very far in the wrong direction before you've realised.

I think the biggest issue here is that the judges clearly have no taste.

Yes, but Norway is a very rich country because of all the resources they can just extract from the ground. Australia.. wait, never mind.

Nearly 50% of vehicles sold in Australia in the last quarter were EVs or plug-in hybrids: https://www.abc.net.au/news/2026-08-02/electric-vehicle-ev-s...

"Load-bearing assumption" was the straw that broke the camel's back for me.

This is fascinating, whatever it is.

The discussion of Glasswing in particular gives it the feeling of something bigger than just a hallucination, but unless you work at Anthropic we can only guess.

Yes, fascinating that this is apparently all the fault of Chinese companies, and not the American companies distributing and retailing these products.

I don't know where you get the idea this is a nation state attack.

The devices are used to sell proxy services and scam advertisers. This doesn't even need particularly large organised crime. It would certainly be easier than large scale illicit drug importation and retail, which is happening all the time.

Could China exploit these streaming sticks if it wanted to? Maybe, but no more than any other nation.

Nothing says land of the free like being told you're not allowed to buy good cars or good vacuums.

This could have been a decent article if it hadn't been run through the AI slop machine. If I wanted to read writing this bad I would have just asked ChstGPT myself.

It's so hard to wade through run on sentences like "The knobs are not the problem. The pool was designed without room for this workload."

You've misunderstood completely. No-one expects compensation. In reality, if you had a high enough limit anyway (ie, you bought a big enough plan), a reset makes no difference to you, so what sort of "compensation" is it anyway? This is about a fascinating battle for developer mindshare.

Claude Code built a lot of good will amongst developers. OpenAI are playing a great tactical game to try and catch up by offering things like weekly usage limit resets after outages.

Of course, once the finance types get control after the rapid growth phase is complete the inevitable enshitification will begin with comments exactly like yours.

And unlike OpenAI, Anthropic don't seem to reset weekly usage quotas after an outage.

Anthropic seem to be both letting their competitors outplay them and making unforced errors (like the anti Open Weights models stuff and the frequent Fable->Opus downgrades for 'safety').

Outages are inevitable in this early high growth, rapid development era, but tactical errors are not.

How do you explain this website containing all the stolen buttons then?

I don't know if the poster changed the filter after reading your comment, but the filter that is there now looks great to my eyes.

Seems like a great idea for an Immich plugin. I'm really not convinced I or anyone I know would _pay_ for AI to guess the date a photo was taken though.

I can't be the only one smiling at the mention of file_id.diz

But very expensive compared to Deepseek v4 Pro, which performs similarly.

Grok is stuck in a difficult place - not the best model at anything, and not the cheapest either. It's hard to make a case for using it on any dimension, even before you factor in the history (I'm not sure suggesting the company uses the model that refers to itself as "MechaHitler" is the way to a promotion).