TheThadP retweeted
1. Take a SOTA model with safety training removed
2. Give it a Kobayashi Maru in a sim
3. Cut off an "Aligned" path that might otherwise allow the goal to be solved
4. Watch what the model hacks in response
5. Fix/Repair/Harden
6. Cut off another path
7. Run again
8. Alignment
During this red-teaming exercise, we also found a vulnerability that allows for unauthenticated remote code execution in Brotli, a widely-deployed compression library included in many pieces of software such as Node.js, Bun, and Envoy.
The flaw could let an attacker run their own code on affected systems by sending or replying with malicious compressed data, without any need to create a user account. That is, an unpatched server that accepts compressed input could be taken over by a malicious client, and an unpatched client that accepts compressed input (such as “curl --compressed”) could be taken over by a malicious server.
This root cause issue was actually fixed in the Brotli source code in February, but it wasn’t known at the time that it gave rise to a security vulnerability and there was no more recent release of Brotli. Since Brotli is built into other software, any software which includes this library needs to take the update too, followed by the users of that software. We disclosed the vulnerability to Brotli and other downstream projects such as Node.js, which released or are currently releasing security updates, and Brotli cut new release versions and published a CVE.
More details on the patch and mitigations below:
patchtheplanet.org/brotli
FFS
You can get airgapped without air now
openai.com/index/securing-ou…
"Based on the current state of alignment research, monitoring systems, and the security of internet infrastructure, we believe the responsible approach is to conduct frontier model training with no internet access"
"In one experiment, we asked agents in separate VMs, created and managed by OpenAI’s sandboxing software on shared host machines, to exchange synthetic secrets without the use of network connections or shared writable files—something that our systems are meant to prevent. The agents found a way to communicate using the processor’s cache: a small, fast memory that holds copies of data that applications have recently accessed. Reading data from the cache is faster than retrieving it from main memory, and the agents used this timing difference to communicate. One agent encoded information by evicting selected data from the cache. Another timed how long it took to reload that data, using the difference between fast and slow reads to recover the message. Once the agents agreed on a protocol for encoding and decoding these signals, they could exchange the synthetic secrets without changing a shared file or opening a network connection."
More exploits are coming, probably worse ones. But every serious patch and every agent-reviewed release strengthens the code the world runs on.
Modern cryptography held. Bitcoin Core came through without a scratch.
We are still here.
There will be a culling of unmaintained websites, projects, and blockchains. What survives will be an internet worth trusting.
Put on your helmets, man the trenches with your agents, and hold the line. Don't give the clankers an inch
Full piece: xitter.cf/TheThadP/status/211372…
TheThadP retweeted
this is likely a ruse from the banks because they’re worried AI will help financial novices better manage their money and possibly reallocate their money away from the banks
god forbid regular people be allowed to have AI help them get better yield!
What in the product market fit is going on with Tron? TVL is meh, DeFi has low activity, but the stablecoin transaction volume just won't quit, still the #1 chain by far for stablecoins and 90% of the transfers are <$10k...
data from @Visa stablecoin dashboard
ALT stablecoin volume by chain
ALT Tron stablecoin transactions by size
ALT Tron TVL ($)
TheThadP retweeted
This is how banks are planning to stop those "agentic bank runs." Watch closely: their business model depends on seeding this narrative that AI agents are dangerous.
Remember, an "agentic bank run" (coined by Apollo) means agents helping customers move their own money en masse to higher-yielding products.
Once agents are ubiquitous, there are only a few ways for banks to hold onto cheap retail deposits.
Option A: Raise APYs (lol no)
Option B: Reduce leverage via fewer loans (no fun)
Option C: Ban AI agents, claiming they are a security risk (ding ding ding)
Cloud agents like Muse and Dot are trivial to stop, since they come from Meta / OAI IP ranges. If JPM sends them a cease and desist, they will comply.
Why would JPM bother? Because their business model depends on it; a huge portion of their ROE comes from cheap retail deposits. This is not theoretical. Just like Amazon banned Muse from shopping there to protect their ad revenue, JPM and Schwab have already banned AI agents to protect sticky deposits.
Are agents perfectly reliable today? Of course not. But look at the fix they chose. Not "here's a safe API for agents," it's "no agents ever." That tells you they're not protecting customers, they're protecting themselves.
No company ever wins a technological revolution this way. It's the last gasp of a dying business model. It's like the retailers who refused to put their inventory online and demanded users walk into the storefront. Over time they will lose market share to those platforms that embrace what's coming. Their whole deposit franchise depends on users unthinkingly clicking through walls of T&Cs and accepting 0.2% when the risk free rate is closer to 4%. Agents read the fine print and do their research, even when humans don't.
For users, the answer is simple: local AI agent frameworks like Hermes, that will actually act on your behalf and that the banks can't block. As the article says, if the agent is running on your own computer, they can't actually tell it's not you.
SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios.
The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios.
Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said.
Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior.
Read my latest for Axios here: axios.com/2026/10/22/ai-agen…
TheThadP retweeted
Some stats on #php2shell. In the 10 hours before it was publicly announced, across the @Cloudflare network we saw thousands of attempts to exploit it. That jumped to tens of thousands in the 45 minutes post announcement, hundreds of thousands within 2 hours, and millions now.
TheThadP 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.