ustas.eth

Leeches to improve testosterone and sexual function. Will look into this and see if worth doing an experiment.

The 3 months comes from the same self-reported timeline as the $500M headline , no Immunefi submission ID, no screenshots, no case reference. The patch commit is public. Anyone technical could reverse-engineer this narrative after the fact. Not saying that happened, but the receipts are missing. Also a fresh account with zero history doesn’t organically hit 1M views. That takes early signal boosts from notable accounts. Feels less like a frustrated whitehat and more like a coordinated hit tbh. Injective not commenting yet is actually the responsible move. You don’t rush out a public statement on a security incident without reviewing all the facts. CT wants drama on their timeline, legal and security teams work on a different clock. Give it a few days.

“Holy shit, my AI agents are too good” Here’s the part we don’t tell you 😭

Update: 4 days of normal GPT-5.4 usage, and I’m already down to 25% of the weekly quota RIP

Switched to Codex CLI GPT-5.4 yesterday after my Claude Max x20 subscription hit the weekly limit a few days before reset. The quality and level of thought surprised me right away. It's the first model I can reliably use for system engineering and trust not to plant a ton of footguns along the way. Much better than Opus in that regard. Opus has a lively mind, but it constantly misses the whole picture and then goes into the infamous psychopathy.

I built the first AI that earns its existence, self-improves, and replicates without a human wrote about the technology that finally gives AI write access to the world, The Automaton, and the new web for exponential sovereign AIs WEB 4.0: The birth of superintelligent life

Good job! Basically, these are baseline results for the common coding agents, correct? Is the dataset that you've used public? Btw, have you noticed any cheating related to the knowledge cutoff dates? I.e., some of the repositories in the list are from the 2023-2024 contests, the modern LLM cutoffs are in 2025 if I'm being correct.