Mansour

Product Manager-AI Native Products, ML

[Mansour @Mansourdam]

Claude Skills is context engineering, not just a feature.

The problem: Claude's default window is ~200K tokens . You can't dump all your docs/standards/workflows into prompts without degradation.

Skills uses progressive disclosure - 3 tiers

[Mansour @Mansourdam]

@mweinbach @mweinbach Hey Max, are you sure these models were trained from scratch and are not distilled from Gemini?

[Mansour @Mansourdam]

@ValerioCapraro @ValerioCapraro Interesting one. Gemini 3.1 Pro gets it wrong at first, but then with the tip it got it. Interestingly, Fable failed even after the tip.

[Mansour @Mansourdam]

@MParakhin @MParakhin I'm really surprised they haven't done this by now. PR reviews are such a pain right now, and I think a Pro/ Deep Think model would help to solve it.

[Mansour @Mansourdam]

@GergelyOrosz @GergelyOrosz could you share some examples of these pushbacks ?!

[Mansour @Mansourdam]

@mweinbach Yeah, its a great ux. Our bet though is multimodal logging, voice-first since it's the lowest-friction way to capture detail, with strong memory handling that personalizes the app over time.

[Mansour @Mansourdam]

@AiBattle_ It's not a good benchmark , Sonnet 4.6 scores higher than Opus 4.6 lol, and there is absolutely no way 3.5 Flash outperforms Opus 4.6.

[Mansour @Mansourdam]

If the claude Mythos rumors are true,that it uses a recurrent-depth/looped transformer with a shared block iterated for latent reasoning,Sakana DiffusionBlocks is directly relevant: it reframes those iterations as diffusion denoising steps, enabling single-pass training instead of K-step BPTT.

[hardmaru @hardmaru]

For over a decade, we’ve accepted that end-to-end backprop is the only way to train deep networks. But holding the entire network in memory all at once is why AI training is hitting a resource wall.

We found a new way to break the network into blocks and train them independently. The trick? Treating the network’s forward pass like a diffusion model denoising a signal.

This reinterpretation slashes the memory needed to train deep models. In our #ICLR2026 paper (https://t.co/PK5h0mqQSo), we matched end-to-end performance across ViTs, DiTs, and LLMs. We did this while training just one isolated block at a time.

[Mansour @Mansourdam]

@badlogicgames Smaller models like Kokoro-82M are already working surprisingly well for TTS
Check this out: https://t.co/wNuIT4F6og

[Mansour @Mansourdam]

@_arohan_ @_arohan_ I think most of the chips in Colossus 1 are H100s.

[Mansour @Mansourdam]

@rauchg @rauchg I still remember our conversation in Amsterdam in 2022, right before the ChatGPT moment. You said a lot of legacy SaaS wouldn't survive long term and pointed to Salesforce as a prime example: clunky UI, slow, bloated. Now with AI, we're hopefully getting there.

[Neil Stone @DrNeilStone]

This person runs an "anti racism" organisation in the UK

I kid you not

[Cloudflare Radar @CloudflareRadar]

Iran's Internet shutdown is now in its third day. Traffic volumes remain extremely low.

Follow the latest status at https://t.co/uMVu3O3zue

#IranProtests2026 #IranDigitalBlackout

[Mansour @Mansourdam]

@raizamrtn Exactly,There are many frontend issues; streaming, for instance, is buggy and triggers multiple re-renders. It makes me wonder if anyone is actually QAing this or gathering feedback.

[Mansour @Mansourdam]

@fleetingbits The recent paper was from Google Research, not DeepMind.

[Mansour @Mansourdam]

@karpathy @karpathy, what is your take on diffusion language models?

[Mansour @Mansourdam]

Once again amazing work by @JeremyDanielFox and the team👏

[Mansour @Mansourdam]

Claude Skills is context engineering, not just a feature.

The problem: Claude's default window is ~200K tokens . You can't dump all your docs/standards/workflows into prompts without degradation.

Skills uses progressive disclosure - 3 tiers

[Mansour @Mansourdam]

Limitations: You're trusting Claude's judgment on which skills to load. Skills can execute code, so audit thoroughly.
The pattern - agents with filesystem access managing their own context through selective loading - is where agentic architectures are heading.