building @infertrail and memes. someone's using your llm keys and it's not you. i catch that.
Joined February 2014
- Tweets6.7K
- Following706
- Followers1K
- Likes695
Pinned Tweet
Your inference may already be getting resold — and your logs can still look normal.
If you run an inference API, gateway, or AI product and you’ve seen suspicious credential sharing, resale, or unexplained usage, DM me. I’m looking to talk to teams seeing this in production.
There is a growing reseller market built around stolen API keys, shared subscriptions, proxy access, account farms, and piggybacked sessions.
The scary part: much of it can look completely legitimate from the provider side.
Valid credential. Valid account. Normal-looking requests.
Meanwhile, the same access can be passed through brokers, resold across borders, or quietly used by someone the account owner has never heard of.
Most LLM security products focus on what an agent is doing.
The blind spot is who is actually behind the account.
That’s what we’re building at @infertrail: behavioral detection for inference abuse, credential resale, and unauthorized access.
We’re publishing honeypot data, reseller-market research across Russian- and Chinese-language communities, attack patterns, and what we learn building detection for this problem.
Follow @infertrail for the research.
i used to be the person who explained why ai cost what it cost.
then i started seeing usage we couldn't explain.
since then i've seen people steal inference for personal use, for distillation, and for offensive security.
inference is valuable. someone pays for every request. steal access and someone else pays for yours.
but i think the appeal goes beyond free compute. using someone else's account puts distance between you and the activity. the account owner is visible. the person behind the requests can be harder to identify.
my research points to a billion dollars in lost revenue.
i've found sophisticated operations with their own payment systems, gateways, account theft operations and marketing. some have 100+ employees.
shell companies in the uk and the virgin islands. servers in russia and china. operators in the us and elsewhere. some advertise openly on x and other platforms. some have vc funding.
that's where trying to explain an ai bill took me: researching the businesses behind stolen inference.
who is actually using the inference you're paying for?
give me a follow if you want to learn how to secure yours. i'll share practical tips on protecting it from llm hacking, key theft, session theft, subscription piggybacking and more.
Arielle retweeted
1/8 **CRITICAL SECURITY ADVISORY: CVE-1985-PING2SHELL**
**Severity:** Critical (CVSS 10.0)
**Affected Systems:** UNIX, Windows (all iterations dating back to 1985), and virtually every networked device since the dawn of TCP/IP.
Our incident response and threat intelligence
while you’re patching your s, check who’s using your API keys. i’m building Infertrail to detect stolen inference and credential abuse.
our website uses HTML so this is a difficult time for us too
infertrail.com
This is unfortunately the future for western models. If you want to do anything interesting in the edges of tech you will be forced to either self host or use abliterated model providers
So I guess Mythos isn’t the beast Anthropic kept saying it was at infosec? At least not in the hands of normal SWEs
wait, PHP was actually scanned by Anthropic’s Claude Mythos earlier this year. they also had access to OpenAI’s Daybreak Red for validating exploits.
the PHP Foundation revealed this JUST yesterday.
keeping frontier cyber models in the hands of a selected few doesn’t seem to be working well for the internet.
thephp.foundation/blog/2026/…
Time for someone to redirect mail.google.com to aol.com
AAAAWWOOOOOOOGAAAAAAAAA
A high severity vulnerability in "the most widely deployed DNS software on the internet" just dropped
kb.isc.org/docs/cve-2026-943…
