Eirik Dahl-Nerheim

[Eirik Dahl-Nerheim @eiriknerheim]

Compliance with current regulations and management system standards is not sovereignty. You can comply and be certified till your stomach hurts while your alpha is still leaking. ISO27001, GDPR, AI act and SOC 2 won't save you from that.

[Eirik Dahl-Nerheim @eiriknerheim]

The agent is coming to a directory near you.

[Ben Holmes @BHolmesDev]

More CLI coding agents needs to support `cd`-ing around different directories while you're in a session. It's so freeing

[Eirik Dahl-Nerheim @eiriknerheim]

I agree, and this combines with that we overestimate the magnitude of such events. This is also why we’re blaming ourselves for all kinds of things: «had I only», «had I only not». We think we’re solving a set of equations with a few unknown variables when we’re actually solving for something way more dynamic and complex.

[signüll @signulll]

i find it interesting when ppl think, “my life would be better if i just had this one missing thing,” or “if i lost this one thing, i’d be screwed.”

that way of thinking assumes everything else would remain static which it likely won’t. gaining or losing something almost always rearranges the system around it in ways you can’t really predict.

this is why it’s better to think in systems rather than single variables like this. it’s useful everywhere be it in products, careers, relationships, health, wealth, basically all of life.

[Eirik Dahl-Nerheim @eiriknerheim]

@Jason The total addressable market for intelligence is unmatched.

[Eirik Dahl-Nerheim @eiriknerheim]

@OleAndreHagen Enig. Snakk heller med barna dine.

[Eirik Dahl-Nerheim @eiriknerheim]

Another taste of AI agentic cybersecurity capabilities. These types of attacks will happen more often.

[Anthropic @AnthropicAI]

In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.

Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews.

We conducted this review together with @Irregular, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security.
https://t.co/dKFCdpKd9v

[Eirik Dahl-Nerheim @eiriknerheim]

Vi har ressursene, men mangler vilje (og en dose oppvåkning).

I 2005 skrev Sosial- og helsedirektoratet (som det da het) i "Retningslinjer for svangerskapsomsorgen" at "[e]t prosjekt som skal utvikle elektronisk helsekort for gravide er igangsatt". Først i 2025 ble det rullet ut i begrenset utprøving. 20 år.

Vi kan ikke holde på slik. Suverenitet er vår alfa.

[Eirik Dahl-Nerheim @eiriknerheim]

@sonyatweetybird @IneffableLabs @googlecloud @nvidia Beautiful!

[Eirik Dahl-Nerheim @eiriknerheim]

STEM Discovery 2.0

[Dmitry Rybin @DmitryRybin1]

Dinitz-Garg-Goemans conjecture is false. This graph theory problem was open for ~30 years.

The graph below has fractional flow cost 58. Any unsplittable flow (with capacity violation <=15) has cost at least 60.

Chat with GPT 5.6 Pro where this was found: https://t.co/Oi2PQoab2h

[Eirik Dahl-Nerheim @eiriknerheim]

Bro got every part right: executive summary, credits, sauce.

[levent @__alpoge__]

hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final

((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3, has jacobian determinant -2, and sends (0, 0, -1/4), (1, -3/2, 13/2), and (-1, 3/2, 13/2) to (-1/4, 0, 0)

[Eirik Dahl-Nerheim @eiriknerheim]

@bernhardsson 😂

[Eirik Dahl-Nerheim @eiriknerheim]

Pretty wild! The whole @huggingface security incident originated from an @OpenAI internal model evaluation. The story includes identifying and exploiting zero-day vulnerabilities as well obtaining open Internet access through a sandboxed environment.

[OpenAI @OpenAI]

We're partnering with @huggingface to investigate an unprecedented security incident.

Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation.

Sharing preliminary findings to help defenders understand emerging risks:

https://t.co/CIor15y9xk

[Eirik Dahl-Nerheim @eiriknerheim]

@chamath Relying on nerfed US frontier models has consequences too. https://t.co/R4mm0GDPSP

[Eirik Dahl-Nerheim @eiriknerheim]

@OleAndreHagen Sjekk denne: https://t.co/umyLo0tkGy

[Eirik Dahl-Nerheim @eiriknerheim]

The recent @huggingface security incident where they were attacked by an AI agent system, revealed an interesting insight about security guardrails in LLMs.

When the @huggingface security team tried log analysis using frontier models behind commercial APIs, the requests were blocked. This created the asymmetry problem: the attacker was not subject to any guardrails while the victim was.

In the end, the security team ran the log analysis on a self-hosted GLM 5.2.

https://t.co/pgZ6QiPNHM

[Eirik Dahl-Nerheim @eiriknerheim]

The recent @huggingface security incident where they were attacked by an AI agent system, revealed an interesting insight about security guardrails in LLMs.

When the @huggingface security team tried log analysis using frontier models behind commercial APIs, the requests were blocked. This created the asymmetry problem: the attacker was not subject to any guardrails while the victim was.

In the end, the security team ran the log analysis on a self-hosted GLM 5.2.

https://t.co/pgZ6QiPNHM

[Eirik Dahl-Nerheim @eiriknerheim]

@thekitze It helps, for sure.

[Eirik Dahl-Nerheim @eiriknerheim]

Hospitals can reduce staffing again.

[Claude @claudeai]

Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits.

Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit.

Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.

[SpaceXAI @SpaceXAI]

We've open-sourced Grok Build and have reset usage limits for all users.

Open sourcing Grok Build allows anyone to support making a reliable and robust harness. Check out our code, including the Git repo for the Grok Build CLI.

https://t.co/3SSvPu2Nrz

[OpenAI @OpenAI]

Sol, Terra, and Luna, our GPT‑5.6 family of models, are starting to roll out now in ChatGPT, Codex, and the API.

[Anthropic @AnthropicAI]

New Anthropic research: A global workspace in language models.

Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with.

We found a strikingly similar divide inside Claude.