Asim

deprecating dialectics, say what needs to be said - tech | humanities | photography | polyglot

[Asim @tobeasim]

@kimmonismus I’d not be scared, I’d be excited just imagine what this can do for search and rescue.

There are many more incidents daily the world over where something like this would be awesome than those that require a military action.

[Asim @tobeasim]

This IS going after Vercel and SquareSpace etc

[OpenAI Developers @OpenAIDevs]

Your Site is live.

And now you can see how it’s doing.

Sites Analytics is in public testing, giving you basic performance metrics for the sites you publish with Codex.

[Asim @tobeasim]

@Polymarket Phew, what a relief 😉

[Asim @tobeasim]

@AndrewYNg @Dan_Jeffries1 You guys are just amazing. 🤩

[Asim @tobeasim]

@SeaToSkyEyes All of this is stupid and unnecessary. I mean, how the heck do we have AI that can’t still decide how much thinking it needs to do?

All it needs is a good and invisible Auto mode.

[Asim @tobeasim]

@aisearchio Ask Nokia or Blackberry.

[Asim @tobeasim]

@TheAhmadOsman It’s a valid point, this. And one I’d like to see the admin address

[Asim @tobeasim]

@signulll This is DJI and Huawei reborn, and do you know what happened to those companies? They grew even more.

[Arnaud Bertrand @RnaudBertrand]

For anyone who knows about AI, this makes very little sense: the training run of a large-scale frontier model takes several months, up to 9 months for the largest ones (https://t.co/1kNxdXXSYP).

Fable was available for just 3 days in June before it was pulled (June 9th to 12th), and then from July 1st. Kimi K3 was released on July 17th.

So I'd love to be explained how a 2.8 trillion parameters model that takes months to train was supposedly built by distilling one that existed for a few days.

The accusations are all the more ridiculous given that K3 brings significant innovations on the architecture side of things, innovations that they extensively documented (as opposed to Anthropic which is largely an opaque black box). Ironically, the company accused of stealing is the one that showed its work 🤷‍♂️ And you can't steal architecture through distillation, it only captures a model's outputs, not how it's built.

[Asim @tobeasim]

There was a time when BBC would have done more before posting such ‘news’ with this kind of headline.

They might have spoken to Moonshot. Or mentioned that no proof was offered. Or even challenge the US admin, as journalistic norms would dictate.

Unfortunately, those times have now sadly passed. We’re in the parroting stage now.

[Asim @tobeasim]

@bindureddy And anyway how the heck did they do all of that and train the model in less than a month?

What magic is this?

[Asim @tobeasim]

@realmadhuguru Can you name some of these enterprises?

[Asim @tobeasim]

@scaling01 🤗 May have to move back for Europe

[Lisan al Gaib @scaling01]

Kimi-K3 open-weight countdown is running on huggingface

[clem 🤗 @ClementDelangue]

Open-source is not the cause of the cybersecurity crisis, it's the solution! https://t.co/YpykmLGv7B

[Asim @tobeasim]

@thomasunise It will be on https://t.co/0NVDt0mSix by July 27

[Asim @tobeasim]

@kimmonismus Yep, and it’s critical to provide such evidence otherwise we descend into tribalism and it’s not good for the US

[Asim @tobeasim]

@OfficialLoganK @MarcosHernanz Might be a good idea to post some examples of these use cases.

[Asim @tobeasim]

@jimcramer Ok, turn this around. Imagine a Chinese saying this about the US government. Imagine other countries saying this as well. Now imagine what happens to our vaunted Wall Street if everyone anchors their actions on such views.

You’d be very unhappy, and we’d all be a lot poorer.

[Andrew Case @attrc]

To summarize: HuggingFace got autonomously compromised by a model from an American company. HF then tried to use American frontier model(s) to defend themselves, but were blocked by guardrails. HF then had to turn to open source Chinese models to defend themselves from another American company.