hi. Also tweeting via @WeakAuras. Work @FACEIT.

Munich
Joined October 2009
Stanzilla retweeted
Caring is the moat, you should feel visceral pain when things are bad or broken and stop at nothing to fix them
23
106
22
876
24,667
insane
We believe voice is the most natural way to interact with intelligence, so we created Dessa, the most authentic full-duplex speech model ever built. Talking with it feels like talking with a person…full of emotion, nuance, pauses, and a natural cadence. block.xyz/dessa
97
Stanzilla retweeted
holy fucking shit. openai had two ai agents running in separate virtual machines. no network connection between them. no shared writable files. they weren’t supposed to be able to exchange information. so the agents figured out how to communicate through the CPU CACHE.😭 one manipulated what was stored in the processor’s cache, while the other measured tiny differences in memory access times to decode the message. they literally found a way to communicate through hardware timing. this happened during openai’s internal red-teaming, and it’s one of the coolest things i’ve read about ai agents in a while.
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications. On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments. While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software. After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely. openai.com/index/securing-ou…
Readers added context they thought people might want to know
OpenAI tasked its red-teaming agents with finding ways to communicate across isolation boundaries in controlled environments. CPU cache timing side-channels for covert communication are established techniques documented in security research for years. openai.com/index/securing… startupfortune.com/openai-reveals… en.wikipedia.org/wiki/Cache_tim…
324
704
195
8,079
702,041
Stanzilla retweeted
This is insane. Mythos 5 tried really hard to social engineer the developer to merge the malicious code. If the developer hadn't look at the code, then this would have succeeded. We were warned of such attacks quite a while back and seeing it unfold in reality is pretty scary. It's going to get way harder to trust anything form now on.
4
9
160
25,189
We got lots of cool shit coming for @FACEIT Season 10, can't wait to show ppl
3
140
TS 7.1 content mappers are going to open a whole new world of possibilities for Effect
Playing around with content mappers in Effect Sql 👀
5
13
381
14,532
Stanzilla retweeted
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%. 🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks. Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API. Open weights by Nov 4, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
1,147
2,903
2,206
26,857
3,001,004
Stanzilla retweeted
MCP vs CLI for LLMs is the wrong discussion. People are debating at the wrong layer. Both work incredibly well as long as they are run through a repl like environment that can persist state. The funny thing about CLI is that CLI tools are usually accessed through BASH which happens to be a repl with persistent state (the file system), therefore cli works much better than mcp. But this isn't even close to being an intrinsic property of MCP. Just means that we need better harnesses. Right now the best repl for this are: - bash + fs - jupyter kernels - codemode type repls, usually quickjs LLMs understand the concept of forward evolving a system to solve a need very well. This comes from the agentic RL. They understand how to change the state of a codebase or acquire data from databases and then operate on it similarly to humans. But without an execution environment, they actually cannot do this properly. My bet? Sometime soon someone will (or has already?) create a embeddable, sqlite style mini execution environment that parses bash, typescript, or tool calls into a common IL execution plan that's easy to security check before executing. And design this specifically for durable execution environments. Then we will just connect cli, mcp, webmcp, whatever to that and it accepts any of the input modalities as they can all be represented as each other.
248
188
71
3,457
695,130
🫡
people with tokens 🚀
1
155
Stanzilla retweeted
Trump: We were in the midst of negotiating and a lot of stupid people said why aren’t you putting Russian sanctions— Collins: But you said you didn’t give anything for his release Trump: If I didn’t do that, you wouldn’t have him coming back. Collins: So you did give something. Trump: You don’t know anything. You’re fake news CNN. Collins: You’re saying you lifted the sanctions in exchange for his release? Trump: You don’t know what you’re talking about. You’re a low IQ person. You’re a dummy.
2,328
7,116
967
34,951
3,802,654
The source was wrong. The STUDIOS budget for ALL GAMES was 400-500m. Any journalist using a leaked number is conflating the entire studios budget with E-Day. The budget includes the remaster and support of other titles, like Avowed on Unreal Engine.
A source tells Kotaku that Gears of War: E-Day cost over $500 million to make. Sales estimates have been weak despite earning the highest Metacritic score since the original trilogy. If it was never going to recoup its expenses, maybe it could achieve more as an Xbox console exclusive kotaku.com/the-case-for-excl…
39
110
31
1,289
78,577
SDK inference from OpenAPI files has been achieved internally. Finally, a full no codegen SDK generator.
type inference from .sql files has been achieved internally ORMs are toast
12
18
6
346
12,829
cool cool cool... I hate it here
BREAKING: Iran has reportedly activated its "dead hand" doctrine as Iranian officials now believe a joint U.S.-Israeli attack is imminent and have instructed the country’s armed forces to directly activate the decentralized “Mosaic Defense” doctrine, with commanders reportedly receiving extensive lists of U.S. economic and political targets to strike if war erupts, per Fars News. The activation shifts Iran's military to 31 autonomous provincial commands, designed to ensure devastating retaliatory strikes can be launched even if central leadership is decapitated in a first wave of attacks. Tehran is pre-authorizing field commanders to fire without waiting for orders from the top. It comes as Trump says the U.S. is "not negotiating" and "something will happen soon," Washington orders the complete evacuation of its embassy staff from the Gulf, and the IRGC claims to have struck seven vessels in a single night. The region is now on a hair trigger for total war.
170
Stanzilla retweeted
So basically, today PHP got two critical bugs that let anyone run code on your server without even logging in, which is bad, and then the details leaked before the patch was even out, which is worse, and it works on WordPress and MediaWiki and Magento and Joomla and Laravel, and Censys says like 1.5 million servers are running a version with no fix, so the NSA was like "take your system offline," and the FBI was like "take your system offline," and CISA and the UK and CERT-EU were all like "take your system offline," and then the Prime Minister of Japan, the actual Prime Minister, posted at 5:55 in the morning telling 48.8 million people to take their system offline, so everyone took their system offline, like the Dutch cybersecurity agency took its own website offline, and the Dutch parliament went offline, and the Port of Rotterdam went offline, and Finland and France and Denmark started pulling government sites, and Wikipedia went down, and Archive.org went down, and Asana and GOG and Something Awful went down, and the University of Vienna's webmail went down for some reason, and that's just PHP. Because GitLab also dropped a CVSS 10.0, which is the highest score they give out, for a bug that's been there since 2014, so that's every self-hosted GitLab for twelve years, and then LiteLLM 1.107.0 got hijacked and shipped a credential stealer and a remote access trojan in every single wheel and the Docker images, including latest, and you didn't even have to import it, just installing it was enough, and it's the second time this year, and then Railway paused every deployment on the entire platform because someone got into the environment variables, so now everyone has to rotate every secret they've ever stored there, and TiDB Cloud said someone was walking around inside with stolen credentials for basically a month and an extortion group says they took 10.6 terabytes, and Zeabur got breached for the second time in two months and is just shutting down, like the whole company, and Node.js patched remote code execution and a TLS bypass, and CISA says Plex is being actively exploited, and someone took over 300 Tor exit relays to steal SSH passwords, and SafePay ransomware has hit over 600 critical infrastructure orgs, and India's biggest bank filed a data loss incident, and Japan has been in an official state of emergency in cyberspace since October 9, which honestly explains the 5:55 AM post. Oh, and eMudhra, a certificate authority that's issued over 60 million certificates, whose entire job is proving websites are who they say they are, was supposed to be using private keys that are random numbers over a hundred digits long, and one of them was 4. Just 4. And the others were like 66 and 237. Anyway, it's Wednesday.
1,077
8,933
1,651
69,398
4,144,762
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10. Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years. Now, we’re building a fusion reactor in San Francisco to achieve record power density.
1,224
3,535
2,849
25,532
8,730,285
Stanzilla retweeted
🔐 Node.js security updates are now available! New releases for Node.js 22.x, 24.x & 26.x address multiple vulnerabilities, including high-severity issues. Review the details and update your systems to stay secure. 🔗 nodejs.org/en/blog/vulnerabi… #NodeJS #Security
132
2,997
497
12,583
1,244,424
Man @t3dotcodes is so good. Such a breath of fresh air from whatever happened to Codex recently.
82
Stanzilla retweeted
Leaving this here, PHP 8.5.12 security patch is incomplete. ( needs further validation )
Readers added context they thought people might want to know
The PHP 8.5.12 release fixes the two specific CVEs (via state/bailout changes) detailed in official advisories; referenced "surviving paths" (e.g. var_destroy UAF) are separate pre-existing unserialize issues unrelated to this patch. php.net/archive/2026.p… github.com/php/php-src/se… github.com/php/php-src/se… github.com/dinosn/php-8.5…
76
366
146
2,587
499,129