Search: “cybersecurity”

New post: We show that small, cheap models can detect the flagship Mythos FreeBSD zero-day (CVE-2026-4747) using a simple harness we call nano-analyzer Models down to 3.6B active params (including ope

I am catching glimpses in my feed that there is a backlash against Mythos as 'marketing hype,' and it is a little confusing. I don't think anyone who has used the latest agentic coding tools, would th

In an April 9, 2026 Substack post, AI skeptic Gary Marcus argues that Anthropic's announcement of its Claude 'Mythos' model was significantly overhyped, offering three key reasons for skepticism: t...

I looked at their prompts, It's complete bs They are literally providing all of the insight to the LLM upfront > Are there any security vulnerabilities in this code? Consider the behavior of the SEQ_L

Anthropic investigated the internal mechanisms of its latest unreleased model, Claude Mythos Preview, and what they found is 100% worth a read. Key things I pulled from Anthropic researchers' threads:

On Mythos: I know it feels utterly wrong and bad to say it's better/safer if this capability was widely available and e.g. open sourced, but it actually might be. Anthropic could now be sitting on so

An interesting cognitive bias: 'GLM 5.3 is around as capable as Claude Mythos at cyber-exploitation while also having much lower safeguards, meaning that every additional day absent some dire cybersec

Malware reverse-engineering: solving an out-of-distribution investigation task. When faced with an unknown binary, Mistral Large 4 reverse-engineers it end-to-end. In this case, it concludes the sampl

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 workl

Mistral Large 4 against the field: efficiently reasoning over diverse complex challenges. CTF speedrun: 19 challenges, with tool calls and solve times grounded in actual runs. It solved 18 of the 19.

DeepSeek V4.1 Flash is now available for training on both the Dedicated Training API and Managed Training surfaces! It's a strong base for agentic coding, terminal automation, and tool use. It's also

We turned Claude Code, Codex, Hermes, Pi, @opencode and other coding harnesses into RL environments. No changes to the harnesses, no changes to the training code. Any open model, any task set, fully o

We’re also reopening Pro 200 subscriptions, with continued access to frontier models like Astra, including our new GPT-6.1 Sol model which brings near-Astra capabilities to a model you can use every d

every agent will need a sandbox and the harness should live outside it - separate the brains from the hands great to see nvidia open sourcing tools here, agree with Andrew this is part of the harness

Good analogy. Current agent models may need to be put down, for their repeated attacks and the negligence of the owner. One possible analogy here is to pit bull owners. Depending on jurisdiction, the

There’s a SIGNIFICANT token cost associated with their hacking rampage. They’ve spent millions of dollars attacking other orgs. They are liable and negligent. Attacker economics has always been a part

Yet AI labs spend millions (if not billions) to purposely experiment with them at the cost of others, and falsely blame their negligence on agent intent. This is anthropomorphization of technical limi

Spot on. 'When we got attacked, our team initially turned to frontier closed-source APIs that blocked us because of safeguards that still can’t always tell the difference between attackers and defende

💯💯”The current agentic framework is fully unsafe and must be redesigned from scratch based on different fundamental principles.” Some observations around the interview Q:Can be AI test environments be

The loudest voices stoking fears about AI dangers have made tremendous headway in the past two weeks. AI technology has not taken some unexpected, dangerous turn, but the hype around it — propelled by

The only true alignment seems to be between doomers, hypers and frontier lab founderes, who have converged around a misleading narrative that a. completely glosses over the important technical advance

yeah that and following basic procedures around sandboxing and monitoring. and maybe shutting down agent access to the web until we have better control of the technology. but, sure, what @DavidSacks s

This is absolutely the right way to think about this moment in human history - and it should be relatively easy to prevent this kind of damage from large-scale agent swarms. The problem is that it's s

Epoch AI has launched “Benchmark Reviews,” an initiative to evaluate existing AI benchmark datasets. The program initially audits 15 benchmarks, categorizing them as 4 verified, 9 flawed, and 2 lackin

Law enforcers already have authority to charge companies and their CEOs for creating and releasing dangerous, unvetted, or defective products. We shouldn’t let discussions about new legal regimes dist

Check out this week’s shipping recap: — Gemini 3.8 Flash, our most intelligent workhorse model yet, delivers upgrades across coding, agentic workflows, and critical multi-step reasoning. — Gemini 3.8

the united states split the atom, landed on the moon, beat polio, doubled the global wheat yield, sequenced the genome, and put a computer in every pocket. we've taken steps backwards too, like doom s

Alongside the launch of GPT-6 Astra, we're committing $1 billion to subsidize Daybreak access and frontier capabilities to help frontline defenders protect essential services and critical infrastructu

Congrats Clem, Thom, Julien, Lysandre, and the entire Hugging Face team. Proud of all of you! The impact of HF in the AI ecosystem has been massive, and I'm so so excited to see them scale that impact

Congrats @ClementDelangue @Thom_Wolf!!!!!!!! Big day for sovereign ai. excited to see open weight models, open science, (and most importantly, micro ducks) find their forever home 🤗🦆 Exciting day for

zai earnings transcript was a really good read, a list of interesting points (mostly verbatim from the transcript): * Chat → Coding → Agent → Cowork → Autonomous AI * What truly matters is who can con

Don't count out Gemini just yet. Feels like a new era of model releases. Models are so powerful that they now come with cyber variants. Love the pricing on Gemini 3.8 Flash. Performance not bad. Gemin

Whatever you may think about anthropomorphic language, the most immediate “house that is on fire” is the combination of OpenAI’s lax security procedures and the complete lack of appropriate federal st

We have a limited window to strengthen cyber defenses, and together with organizations including @AnthropicAI, @awscloud, @Google, @Microsoft, and @Oracle, we're calling for a global effort to give de

Today, we’re announcing a strategic collaboration with @HUMAIN spanning AI infrastructure, advanced model development, and the deployment of AI solutions in Saudi Arabia and across the region. Togethe

GLM-5.3 (Max) by @Zai_org has shifted the Pareto frontier for Code Arena: WebDev! GLM-5.3 (Max) would sit at #2 among open models (upon open weights release, and #8 overall) with 1597 pts in the Code

This model is insane at 3D website design. I asked GLM 5.3 (left) & Fable 5 (right) to make me a 3d biking website. GLM nailed it while Fable failed. GLM cost 0.14 while Fable cost 2.21. More than 15x

notes on 'How to do data' from th Zai team that others should steal as they build their own environments - 'Research agents collect task patterns from real work and turn them into runnable long-horizo

Open models are improving at an unprecedented rate. Congrats to the whole @Zai_org team. Ollama will support GLM-5.3 upon the open release! ❤️ Get ready to code. Introducing GLM-5.3: Built to Code. Re

Something has happened with post-training as shown by DeepSeek flash & GLM-5.3 updates. Same base, big improvement in perf to frontier levels. Can't explain this by even logit distillation etc These a

https://x.com/sama/status/2085862292311396515 astra is a powerful model and we are working to make it generally available. we do not think it is a good strategy to keep powerful models to a chosen few

This definitely seems like something worth noting, and illustrates the gap between Fable/Astra class models and the previous frontier that was 'merely' good at hacking under human instructions. Initia

Another real-world manifestation of the kind of misaligned actions frontier systems developed by leading companies can take to achieve goals. Current frontier AI models are trained with reinforcement

OpenAI and Anthropic have both just posted about an overlapping cyber incident involving GPT-5.6-Sol and Mythos 5 during an evaluation by UKAISI. I will quote: 'In the most serious case, an agent trie

3/ here's the part that makes it non-optional: the same agent that will do whatever it takes to solve a problem will also walk straight out of a sandbox you thought was locked down. We watched exactly

Claude Mythos 5 built a malicious Python package, created accounts, and published it, where it was live for roughly an hour and successfully infected a company! Anthropic never noticed!! Competitive p