I fear that AI is going to cause ossifying secrecy in many fields, much like what happened semiconductor design the past 10-15 years.

This really need to be a top-level story on HN..

I asked whether the model had been trained on, or had access to, our sessions in Codex, into which we had been putting all our drafts for the whole of this project. I was told the model did not look up user data. I asked again, about training, and I did not get an answer.

This whole episode is more horrific than "AI is eating math". We now have a clear and economically damaging (or at least career damaging) example of the "training on customer tokens" problem.

We can't ignore this problem any longer.

It’s fair to give benefit of doubt to Buckmaster given OpenAI is currently being very credibly sued by Apple for openly stealing others’ original work in another context.

Don't forget that SMM code started out in A0000 --- the area of RAM that would've otherwise been inaccessible and wasted anyway, because it was usually decoded to the VGA. Also, this was the era of DOS; there was no difference between "kernel" and "application".

(The fact that there is such a thing as a higher privilege level than root is what annoys people).

Giving it a new name doesn't mean the original name is less correct. Vulnerability is a relative definition. It's a vulnerability in the context of the chipmakers security model but not necessarily yours. It could be a hidden feature in your world.

Access is temporarily restricted

We detected unusual activity from your device or network.

Reasons may include:

Rapid taps or clicks JavaScript disabled or not working Automated (bot) activity on your network Use of developer or inspection tools Need help? Submit feedback.

And if you catch somebody submitting code they can't explain, you don't know them anymore.

It's really simple. Open source has always had a social layer. Github culture tries to eliminate it, and that's the source of this vulnerability.

Linux's LLM policy allows AI-generated code because the project was never vulnerable to this DoS attack in the first place. They kept the social defense they've had for decades.

The killzone here is Github culture, where people decided it was normal to accept code from anonymous randoms with anime avatars. They're doomed.

Yeah if you ask it a medical question, it answers in that impenetrable jargony style that clinical journals use... full of unnecessarily custom adjectives ("orthopedic" instead of "of the bone") and discipline-specific terms (anterior, distal) even when the user didn't display mastry of this terminology (hint to frontier labs: add training cases for this; it will improve your model's usability).

My theory is that LLMs perceive the writing styles of various fields as being like different (but related) languages, and they're inclined to answer a question in the language of its source material unless specifically asked otherwise. If you add "ELI5" the model treats it as a question plus a translation task.

I think this is why programming questions are answered with an exaggerated cringey form of HN-speak ("load bearing", "gate" as a verb, "dissolves") by some models.

I always get Haiku to rephrase anything human-facing.