Lass Najmuldeen

RL Research & infra @ distributed cognition | ex founding MLE @ Ossus (YC W26) | {}

[Lass Najmuldeen @TensorTortoise]

@michaelfertik @distcog @steven_mottaa excited to continue doing great things together

[Michael Fertik @michaelfertik]

It is already clear to me that @distcog and @steven_mottaa will likely be of monumental importance in humanity's journey toward collective superintelligence.

[Lass Najmuldeen @TensorTortoise]

@juri_taresch that's wasup

[Lass Najmuldeen @TensorTortoise]

@michelangelo_mv @distcog 🐐🐐

[Lass Najmuldeen @TensorTortoise]

@leonard_cor @distcog {}

[Lass Najmuldeen @TensorTortoise]

@Paband_O @distcog huge

[Lass Najmuldeen @TensorTortoise]

@FrancescoSacco1 @distcog soon TBD 👀

[Lass Najmuldeen @TensorTortoise]

@distcog we're so onto it

[Lass Najmuldeen @TensorTortoise]

the swarm is coming

[distributed cognition @distcog]

Today distributed cognition exits out of stealth with backing by Verdict Capital.

Our mission is to replicate humanity's collective intelligence.
Read more: https://t.co/PqqgbLXbkc

Most of today's AI power is held in the hands of a few, who can choose what you're allowed to think and what you're permitted to do.

We stand on three substantially different foundational beliefs:
- You shouldn't ask permission for intelligence.
- Superintelligence will be a distributed collective.
- Humans are not meant to be relegated to a permanent underclass.

If we don't distribute intelligence, nobody will.

[distributed cognition @distcog]

Today distributed cognition exits out of stealth with backing by Verdict Capital.

Our mission is to replicate humanity's collective intelligence.
Read more: https://t.co/PqqgbLXbkc

Most of today's AI power is held in the hands of a few, who can choose what you're allowed to think and what you're permitted to do.

We stand on three substantially different foundational beliefs:
- You shouldn't ask permission for intelligence.
- Superintelligence will be a distributed collective.
- Humans are not meant to be relegated to a permanent underclass.

If we don't distribute intelligence, nobody will.

[Lass Najmuldeen @TensorTortoise]

@AhmedkKadry my guy 🤞

[Lass Najmuldeen @TensorTortoise]

Super happy to share that my first first-author paper was accepted at the NeurIPS 2026 FLMSec workshop in Paris!

Depth-based early-exit methods let an LLM stop partway through its layers and generate an answer early to save compute. We asked whether any of those exits are any more robust to jailbreaks than the full model.

We used Balcony’s frozen-backbone design, which adds lightweight exit modules at intermediate layers while leaving the original model unchanged. Because every exit shares the same weights, we could attack each exit independently, optimizing a separate adversarial suffix for each one, and measure the effect of depth alone.

The two models we tested behaved very differently. For LLaMA3-8B, early exits were substantially more robust: attack success at the earliest exit we tested (layer 15) was only 4%, compared with 28% on the full model. For LLaMA2-7B the gap was much smaller. At the same depth it was already roughly as breakable as the full model (36% vs 44%). We also found that deeper exits were easier to optimize against, and that the adversarial suffixes didn’t systematically degrade utility at any depth. So exiting early helped on both models, but how much depends heavily on the backbone.

Huge thanks to Seyedhamidreza Mousavi for collaborating on this!

Paper: https://t.co/HS8Mv8BkIY
Code: https://t.co/LU3071tZDZ

#NeurIPS #LLMs #AISafety #Jailbreaking

[Lass Najmuldeen @TensorTortoise]

@Engmelli1 🤝🤝🤝🙏🙏

[Lass Najmuldeen @TensorTortoise]

@VedantRath90232 Ty twin 🤝

[Lass Najmuldeen @TensorTortoise]

Congratulations, slide your boy some compute pls

[Modal @modal]

Today we're announcing that Modal has crossed $1B in annualized run rate, as well as raised a $750M Series D at a $15.75B valuation, led by @Accel.

Since our Series C, usage has tripled and we’ve grown more than 15x year-over-year.

We are deeply grateful to the developers and companies building on Modal. There's still so much more to build, and we can't think of a better time to be building software.

[Lass Najmuldeen @TensorTortoise]

@FrancoisChauba1 get @GenAI_is_real there ASAP, goat of SGL

[Lass Najmuldeen @TensorTortoise]

1 trillion params just to score 28% on terminal bench. $3.5 billion series D btw

[Mistral AI @MistralAI]

Meet Mistral Large 4, aka Le Chonk.

• 1T parameters, natively multimodal. 49B active.
It is the best open weights model from US or Europe on aggregated benchmarks.
• State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding.
• Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure.
• Available to all via API today. Working with cybersecurity partners privately.

Open weights release end of October.

[Lass Najmuldeen @TensorTortoise]

@Aleph__Alpha @steven_mottaa new model to abliterate

[Lass Najmuldeen @TensorTortoise]

@VedantRath90232 Monstrosity