artwalkx

 AI Architect

[artwalkx @artwalkx]

🥲
The Wandering Singapore 🇸🇬: No more haze! 🌫️

[laixintao @laixintao]

想到一个可以解决雾霾的方法，叫「流浪新加坡」。

[artwalkx @artwalkx]

Now we knew AI really matters.
It turns out AI really can solve Singapore’s haze problem.

[Formula 1 @F1]

Marina Bay, as seen from the air... 😮‍💨👀

#F1 #SingaporeGP

[artwalkx @artwalkx]

Left Beijing years ago. Never imagined I’d get to experience months of haze 🌫️ all over again — in Singapore. 😂

[artwalkx @artwalkx]

@Polymarket @grok is this true?

[artwalkx @artwalkx]

LOL, Gemini is basically the US version of Cici: insanely fast, but kinda dumb.

I miss Gemini 3.0 Pro — it actually felt better than ChatGPT earlier this year.

[CG @cgtwts]

Dario: we need to slow down AI development.

Elon: Dario is right.

Sam Altman: we agree.

Google: we’ve already been slowing down AI development for nearly a year.

[artwalkx @artwalkx]

😂

[artwalkx @artwalkx]

Is it already too late?
Can humanity even hit pause? Even if you pace, how do you ensure everyone else does?

It’s a classic Prisoner’s Dilemma. Coordination is impossible—the only way out is accelerating straight through the singularity.

[Dario Amodei @DarioAmodei]

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.

Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.

You can read the full post here: https://t.co/OGyPb7yaYt

[artwalkx @artwalkx]

Excellent—it seems as though I were simulated.

[Polymarket Traders @PolymarketTrade]

We gave a fly $100 to trade on Polymarket.

It already called today’s pump in the "25 bps increase" on Fed Decision polymarket.

Every second, its next move is decided by 166,700 neurons.

[artwalkx @artwalkx]

Wow, it's much better than mine.

[Breg Grockman @alright_mark]

the fly has completed* the ultimate test: parallel parking

[artwalkx @artwalkx]

But it's still trained on human-generated data and learns to mimic human behavior in our world.

[Naval @naval]

Yesterday's AIs ate the Internet and became a patchwork mediocrity.

Today's AI hunts and feasts upon the Mozarts, Einsteins, and Shakespeares of our age.

To birth a gang of ghostly geniuses that write our code, drive our cars, do our work, and run our world.

[artwalkx @artwalkx]

Same feelings

[Anthropic @AnthropicAI]

Anthropic’s Economics team is sharing a new model of how AI might affect economic growth, jobs, wages, and more by 2030.

Explore the scenarios, tell us what you think will happen, and see how your answers compare to more than 10,000 Americans. https://t.co/AvQlEZNxR0

[artwalkx @artwalkx]

This is incredible — any chance this becomes the default in @darkbloomai soon?

[Pratik Gandhi @pratikg]

BREAKING: @googlegemma just pointed their audience at the leaderboard that made Gemma 4 26B A4B 2x faster on Mac.

Some of those gains came from people pointing Gemini 3.7 at it!

@Google's frontier model, in the hands of strangers, speeding up Google's open-weight model. The rest came from GPT-5.6, Claude, Qwen3 and MiniMax on the open https://t.co/uDZcKSPYOm board at @YukonResearch.

The team that shipped the model showing up to watch is new. That only happens with open weights.

Open frontier research accelerates open intelligence.

[artwalkx @artwalkx]

Darkbloom turns idle Macs into distributed compute nodes for AI inference.

Setup is super quick and painless. Excited to see more open-source models being supported to meet market demand!

[Pratik Gandhi @pratikg]

Folks! Another 100 Macs joined @darkbloomai

That is in the last 14 hours!

Make that 200+ new Macs in 60 hours.

Put your Mac to work 💁‍♂️

[artwalkx @artwalkx]

At first, I thought it was Plants vs. Zombies.

https://t.co/2k8VRUNErA

[artwalkx @artwalkx]

So true 🥲

[Jen Zhu @jenzhuscott]

True story.

[ali @waterloo_intern]

https://t.co/aivc7B4A7Z

[Cerebras @cerebras]

https://t.co/FvKy1StDpQ

[NVIDIA AI @NVIDIAAI]

We took a 30B model and split it in two to write tokens in parallel instead of one at a time.

Introducing Nemotron-Labs-TwoTower: a diffusion language model from NVIDIA Research adapted from Nemotron-3-Nano-30B-A3B. Here’s how it works: one half holds the context, the other writes the tokens, with both reusing the pretrained model instead of training a new one from scratch.

We found it kept 98.7% of the original model’s quality at 2.42× faster generation.

[artwalkx @artwalkx]

So TURE

[Elon Musk @elonmusk]

Giving people agentic AI be like …

[artwalkx @artwalkx]

Does this imply that Qwen 3.5 models with larger sizes are theoretically even better?

[Awni Hannun @awnihannun]

This is a great thread comparing Qwen 3.5 4B to GPT 4o. And an additional datapoint that they are in the same capability league (with some interesting nuances).

Thanks for the independent evaluation @N8Programs 🙏