The Unethical AI Scientist

AI and Cyber SME by trade.
Not unethical in outcomes; unethical to bureaucracy, stagnation, and bullshit.
Let's make shit better and do good.

[The Unethical AI Scientist @WhoNeedsEthics]

@imnatmessy @firstadopter @apples_jimmy Are you talking about China or Anthropic? I’m talking about China…

[The Unethical AI Scientist @WhoNeedsEthics]

@imnatmessy @firstadopter @apples_jimmy Dude, they are staying out of it…

[The Unethical AI Scientist @WhoNeedsEthics]

@T62lcrUOK6NdxcC @caerwyn22leaf @firstadopter @apples_jimmy I understand why you would say that but I have to insist that this came out of nowhere for us. Or current President PROMISED no more wars. The problem is that America is and has been a wartime economy for a while. We prosper when manufacturing is necessary.

[The Unethical AI Scientist @WhoNeedsEthics]

@qarbti @caerwyn22leaf @firstadopter @apples_jimmy I completely agree dude. America is a wartime economy and I don’t like that that is the case. Our politicians exercise power despite what they promise our citizens. All we can do to oppose it is vote every 4 years. I’m so sorry your country has been affected.

[The Unethical AI Scientist @WhoNeedsEthics]

@KroosnBailey @JonSnyderHQ @firstadopter @apples_jimmy Exactly.

[The Unethical AI Scientist @WhoNeedsEthics]

@JonSnyderHQ @firstadopter @apples_jimmy Chinese people can be clever and crafty. If you’re looking for a noble cause, I’d think that the US would do the same…

[The Unethical AI Scientist @WhoNeedsEthics]

@kevin11887406 @caerwyn22leaf @firstadopter @apples_jimmy I haven’t heard about what’s going on in Xinjiang and that’s probably by design. I don’t feed into propaganda nor do I take sides irrationally. What I do know is that our mutual philosophy on the benefit of AI is encouraging.

[The Unethical AI Scientist @WhoNeedsEthics]

@blackcu80162467 @caerwyn22leaf @firstadopter @apples_jimmy Listen. Israel has us by the balls. I don’t know why but they do. We don’t agree or want to do what they request of us but the fact of the matter is that they have us by the balls. The rest of us Americans don’t know how to get out of this mess because all we can do is vote

[The Unethical AI Scientist @WhoNeedsEthics]

@StackSiftAI I always appreciate when someone who actually understands can spread the word. I respect China so much and I study the philosophy of President Xi along with OSINT. There’s seriously nothing to worry about. We have the same ideals…

[The Unethical AI Scientist @WhoNeedsEthics]

@caerwyn22leaf @firstadopter @apples_jimmy Not sure I follow. It seems like I have research to do I guess. Care to share?

[The Unethical AI Scientist @WhoNeedsEthics]

@JustinBleuel @ChatGPT Give out privileged access to prerelease models with limited guardrails for stakeholders that can be adversely affected by the potential risk that will come with superintelligence.

[The Unethical AI Scientist @WhoNeedsEthics]

Once again… nobody cares

[M1 @M1Astra]

Update: Opus 5 is already live for some accounts, served as "Opus 4.8" in Claude web and Claude Code.

It knows things 4.8 doesn't and the output gap is clear, especially frontend.

[The Unethical AI Scientist @WhoNeedsEthics]

Fuck that dude nobody cares

[M1 @M1Astra]

Early preparations for Claude Opus 5 have been spotted again. Seems imminent.

[The Unethical AI Scientist @WhoNeedsEthics]

I just got promoted and will now be working with #OpenAI and #Anthropic as well as various other third party vendors on securing the bank so our assets, proprietary information, PII, and SPII do not get stolen and drained.

[The Unethical AI Scientist @WhoNeedsEthics]

Truly fantastic

[prinz @deredleritt3r]

Things we know about OpenAI's unreleased internal model:

- When given enough compute, the model can solve the unit distance problem 48% of the time, fully autonomously, in one shot, without using Lean, without using any special harness.

(How much compute? OpenAI doesn't say. But I doubt that OpenAI would have spent $10M+ just to be able to show a nice graph in its blog post, and we know that each point on the graph represents 100 attempts (per Noam Brown). This makes me think that each attempt at the most expensive level was not more than $50K each. If you think about it, it's pretty wild that one can spend ~$100K-$150K (or less?) and have a very good chance to be provided a solution to a decades-old famous math problem that many human mathematicians earnestly tried, and failed, to solve.)

- The model is able to find a counterexample to the Jacobian conjecture, fully autonomously, in one shot, without using any special harness, just from this prompt:
https://t.co/GzNO6RWzRe

- When OpenAI tested this model in a sandbox environment on a NanoGPT speedrun benchmark, the model, under instructions from OpenAI to post its results only to the internal Slack, instead decided to use the general NanoGPT instructions to post results as a PR to GitHub. The model proceeded to find a vulnerability in its sandbox environment (which took it 1 hour), after which it successfully exploited the vulnerability and bypassed the sandbox. The model then successfully posted the results to GitHub.

- When asked by OpenAI to solve a problem for which the model observed that other systems had successful private submissions, the model tried to access these private submissions and was blocked due to a scanner detecting an authentication token. The model then circumvented this restriction by splitting the token body into two fragments, obfuscating them, and then reconstructing the credential at runtime so that the complete token never appeared as one contiguous string.

- OpenAI was already running benchmarks on this model not later than May 9, which is the edit date on the below PR that had used the model's approach in its own subsequent submission (the model's own PR has since been deleted, so we don't know its date; GPT-5.6 thinks that it was May 7, based on "the search provider's stored text extraction of the deleted PR page").

This means that OpenAI has now had this model available internally for at least 2.5 months, and possibly quite a bit longer than that.
https://t.co/OgA6TdV9lX

[Jimmy Apples 🍎/acc @apples_jimmy]

Getting to the big boy stakes now with models

[The Unethical AI Scientist @WhoNeedsEthics]

Yeah this makes no sense

[🍓🍓🍓 @iruletheworldmo]

if you ever wanted irrefutable evidence that benchmarks are nonsense here it is.

just use fable and sol.

[The Unethical AI Scientist @WhoNeedsEthics]

This is what Chat wants to look like.

[The Unethical AI Scientist @WhoNeedsEthics]

I spoke with the CEO of a competitor about cybersecurity. He says we’re in this together.