minzi

I pretend to know how computers work. Mostly shitposts. Sometimes security stuff/programming

[minzi @minzicat]

@flpsnd @t3dotcodes @pidotdev https://t.co/nDWuduOIze

[minzi @minzicat]

for those who dont care about breaking ToS; ive made a much nicer solution

`pi install npm:pi-claude-oauth-adapter`

https://t.co/kEsyaxO1RM

[minzi @minzicat]

they turned it off :(

[minzi @minzicat]

okay so it's not just me, is this a bug? def not complaining lol

[minzi @minzicat]

okay so it's not just me, is this a bug? def not complaining lol

[SSH @SSHCodes]

am i the only one just having gpt 6.1 sol dump its reasoning

[SSH @SSHCodes]

am i the only one just having gpt 6.1 sol dump its reasoning

[minzi @minzicat]

I can barely get a few agents to work together let alone 10K, how??

[OpenAI @OpenAI]

This model represents a step-function improvement on many benchmarks, and its training is ongoing.

Our internal model group arrived at the Navier–Stokes solution in 88 hours, using around 10,000 coordinating AI agents.

Throughout the effort, we maintained the strict safeguards—including monitoring and isolation—that we apply to all our frontier evaluations.

[minzi @minzicat]

someone pls fund my codex and claude subscriptions

[minzi @minzicat]

@stolinski If you don’t mind breaking ToS: https://t.co/GHUQnk2jot

[minzi @minzicat]

If you want to test Opus 4.7 in a proper harness like Pi and Claude Pro/Max auth is being weird, use this: https://t.co/kEsyaxO1RM

Install:
`pi install npm:pi-claude-oauth-adapter`

It’s the Anthropic OAuth / Claude Code compatibility adapter we use for Pi.

[minzi @minzicat]

Am I the only one who doesn’t understand the hyper around @bot lol

[minzi @minzicat]

@paoloanzn I’ve had this for months :P

[minzi @minzicat]

for those who dont care about breaking ToS; ive made a much nicer solution

`pi install npm:pi-claude-oauth-adapter`

https://t.co/kEsyaxO1RM

[minzi @minzicat]

@arunoda do lmk if this is a reoccurring thing, would love to stop by and meet more people building agents

[minzi @minzicat]

@robzolkos https://t.co/pEwS9E1BzA

[minzi @minzicat]

for those who dont care about breaking ToS; ive made a much nicer solution

`pi install npm:pi-claude-oauth-adapter`

https://t.co/kEsyaxO1RM

[Rob Zolkos @robzolkos]

Quick demo of a (legal) way of using Claude Code subscriptions in Pi. Also a show n' tell of my little slopchop diff reviewer. Enjoy this sweet, sweet taste of Pi.

[minzi @minzicat]

@robzolkos Yup they do, and I have a pipeline that automatically maps all the telemetry and fingerprinting, so that extension gets very close to the real cc shape without adding the system prompts, and gets around the filtering

Obviously there’s a ban risk; hence the disclaimer

[minzi @minzicat]

https://t.co/GHUQnk2jot

[minzi @minzicat]

If you want to test Opus 4.7 in a proper harness like Pi and Claude Pro/Max auth is being weird, use this: https://t.co/kEsyaxO1RM

Install:
`pi install npm:pi-claude-oauth-adapter`

It’s the Anthropic OAuth / Claude Code compatibility adapter we use for Pi.

[minzi @minzicat]

https://t.co/GHUQnk2jot

[minzi @minzicat]

If you want to test Opus 4.7 in a proper harness like Pi and Claude Pro/Max auth is being weird, use this: https://t.co/kEsyaxO1RM

Install:
`pi install npm:pi-claude-oauth-adapter`

It’s the Anthropic OAuth / Claude Code compatibility adapter we use for Pi.

[minzi @minzicat]

Anthropic has started doing some very desperate filtering to filter out third party clients lol

They do a similar thing with pi (they check for pi docs in your system prompt lol)

Take a look at this if you wanna get around it and use your sub on pi: https://t.co/kEsyaxO1RM

[Om Patel @om_patel5]

THIS GUY LOST $200 IN ONE DAY BECAUSE THE STRING "HERMES.md" WAS IN HIS GIT COMMITS

HERMES.md is a real convention used in AI agent projects. it's a system prompt specification file. not some obscure edge case

he's on claude max 20x at $200 a month. yesterday claude code hit him with "you're out of extra usage" out of nowhere

his dashboard showed 13% weekly usage. 0% current session. 86% of his plan was sitting there untouched

but $200.98 in extra usage already burned through what should have been covered by his subscription

he tried logout & login, different models, fresh installs and nothing worked

anthropic support sent the ai bot (four rounds of the same scripted response). eventually they just gave up on him

so he started binary searching repos and commits manually on his own time until he found the trigger

the string "HERMES.md" in a recent git commit message

uppercase, with the .md extension, anywhere in your commit history

that's it

claude code includes recent commits in its system prompt and something server side flags HERMES.md and quietly routes you off your max plan onto API rate billing

> AGENTS.md? fine
> README.md? fine
> HERMES without .md? fine
> lowercase hermes.md? fine
> uppercase HERMES.md? you're getting charged API rates

he reported it. anthropic support acknowledged the bug three times, called it an "authentication routing issue", thanked him for finding it

then refused to refund the $200

so the man pays $200 a month for max, lost another $200 to a billing bug they confirmed, did anthropic's QA work for free on his weekend, and got a "thank you for your patience" in return

check your commit history before claude code quietly drains your account too

[minzi @minzicat]

@om_patel5 lmfao they’ve started fucking regexing strings to detect third party clients

They do the same thing with pi docs (if it’s in the system prompt)

[Calif @calif_io]

MAD Bugs: All Your Reverse Engineering Tools Are Belong to US

Ghidra, radare2, IDA Pro, and Binary Ninja Sidekick. If your tool doesn't show up here, it's not cool enough. Contact us for a free RCE.

https://t.co/PsCenNMKtI

[minzi @minzicat]

@SIGKITTEN okay I need a cat with the same design on my pi setup now

[minzi @minzicat]

@robzolkos cause you notice a drastic improvement when agents only have exactly what they need and not more, esp skills, even 5.4 gets confused after like 15 skills (one trick is to leave breadcrumbs and not the entire list so they can find it) and define the high level philosophy

[minzi @minzicat]

@robzolkos I feel like that’s way too many skills for most users, is there a way users can configure it /go through a setup wizard with an agent who knows the system so it can configure only the things they need? I’ve been putting together something similar but with just the must haves