golly_ned

For years I've heard about this supposed relation between AI and drug discovery.

I can't tell -- has there been any real progress made? Or is this a supposed application of AI that is getting interest? What about this problem in particular makes it suitable to AI as opposed to any other?

Or is this just a matter of -- "AI can cure cancer and is Good For the World!"

They represent the ~99%+ of income. X is large enough that extremes are represented. Don't be fooled by the extremes.

Go after the achievable P75-P90 that's within your power.

80%+ of the article is not 'save money'. It's nuanced.

I'd add or emphasize being strategic about which roles within an organization are rewarded. These are usually roles with very clearly attributable impact to top-line or bottom-line metrics.

Much enablement, unblocking, capability-based, or risk reduction work is very difficult to interpret for the business decision-makers.

Massive cognitive offloading and possibly deeply net negative in terms of mathematician hours and attention depending on how far this goes.

What proportion have actually been verified by the mathematical community by their own standards of proof? There is dispute in general about what constitutes proof, and which statements qualify as having been proved.

But even formalization doesn't mean the formalization corresponds with the stated problem.

Managers can be even worse. I’ve seen an entire roadmap mostly AI generated. Director+ across the company approving based on it having the right keywords, without nuance or detail.

I’ve seen monthly and quarterly executing results AI generated with plainly wrong factual information.

I worked with someone like this for a few years, and it was the biggest mistake of my career.

He'd talk at me for 15 minutes if I asked him a yes or no question. He'd take over a room to talk and was constantly angry at the mistakes of others or why they're not doing things he wanted. He was incredibly controlling with me in particular. Any mistake I made he would try to call out in the group to shame me. I couldn't even go on vacation without him pointing out my absence as if I'm doing something wrong.

I had my first panic attack in my life about six months after joining the team. I started taking a very powerful and uncommonly used anti-depressant that first year that's especially good for anxiety. I gained 65 pounds from it and stopped being able to sleep naturally without medication.

My career was going fine. More responsibility, more prominence, more success in particularly visible positions. But he knew he was more technically proficient, and especially had a lot more experience with some of the technologies we were using, and became seethingly jealous. In a very tense cross-org, principal-to-principal design review for a document I'd written, he took over the room and ambushed me with tons of criticism. He had a chance to review it internally in our team, but didn't. He waited to try to make me look bad in front of a very powerful audience.

It didn't work. They just moved on from his discussion. He got angry at the rest of the room when the topic moved towards how the teams should work together instead of his the document itself, which he had more criticisms about. During a team retro, someone (not me) anonymously noted that down. He furiously glared at me.

After the big launch which was very successful, our team was celebrating and congratulating each other. Someone gave me lots of praise. He sat there with his mouth agape, incredulous that I was receiving praise. He said he couldn't believe what he was hearing.

My manager then rolled his eyes. He was sick of the fact there was conflict. He thought we two should just resolve it ourselves. The other guy was very useful to my manager when it came to operations review and didn't want to lose him.

All this completely wore me down. Eventually I broke down in tears in front of my manager (a new one) and senior manager. He must have been sent to anger management or something. He stopped outright attacking me.

For years I heard his voice whenever I tried to code or write a technical doc. Only in the past few years did that stop happening. It took away so much joy.

I learned I should have just gone straight to HR and said he was bullying me. I should've gone in with a few examples. Even if it doesn't end up being a credible, established case, and even though HR tends to try to punish people who present themselves as a potential risk to the company, at least it would have given him the right shock to be careful around me.

> Folks get mad at Musk, Zuckerberg, etc... if they're doing something illegal, ok yea that makes sense. If folks are mad about actions that are legal, it makes zero sense to be mad at the actor

You're claiming that legality is the necessary & sufficient condition for. Legal != Good, Right, Just, Moral, etc.

Many things are illegal, but still you'd have to be an immoral, evil asshole to do.

Given the professor's professional interest in cheating-catching tools, I'm suspecting wanting to get good, interesting data motivated a lot of their behavior here, even in ways he might not consciously realize.

> “ You seemed quite nervous”

This kind of thing is weird to call out. Your perception of a candidate’s emotions isn’t relevant. People are wired differently. Some people can bullshit with no nervousness at all. Some people can be highly skilled and not able to practice this away. It’s a matter of wiring.

Calling out actions and behaviors and words, observable reality, is fine. Emotions, no.

+1 to this at big tech.

H1B were no less skilled (and no less paid) than their US-based counterparts.

Personalized search results are not ipso facto unreliable.

Please stop with this psychoanalyis and mind reading with AI and human fear. It’s a thought terminating cliche at this point.

In this case, it’s much simpler and more human. Largely between two humans — buckmaster and Bubeck. The interesting question is what the role of contribution and credit for research in the ai world.

The capabilities of AI aren’t even in question in this case.

Very strangely, they said something directly contradictory. initially that it was impossible to rule out whether bucmkaster’s conversations went into training data. Now they claim the opposite with full confidence.

Where did OpenAI say this?

And it still leaves open the question of the prompt itself, which can just as easily encode information about the same knowledge.

At the very least, a company shrugging and saying it’s impossible to know whether academic plagiarism had occurred is a claim that needs to be justified, not taken at face value.

And even if so, it should be on the company to design systems to avoid academic plagiarism and offer the right transparency. It shouldn’t suffice to say “we don’t know what went into the model, when, or how” —- that’s a solvable problem that an accountable company can satisfy.

To add: with a requirement that he rewrite the proof to credit OpenAI.

The (unprovable, yes, without OpenAI being willingly transparent) argument is that openAI constructed a prompt to scoop them using some inside knowledge about the approach, which they allude to in the announcement.

In the transcripts, Brubeck is very cagey and evasive about the prompt, when it was supplied, and its contents.

This is wrongheaded in so many ways.

To start -- how many of those 3,000 engineers do you think work on the mobile app? And even among those who work on the app -- do you think they're all working on user quality-of-life and just can't get it right? And it's a $175BB company -- do you think it's worth that much based on the quality of its mobile app?

When it first released, the criticism that it's mostly empty procedural generation what what got currency and stuck to the no man sky name.

That still sticks around as people who've never played it comment & upvote the same things.

I've never played it. Just witnessing the cargo culting. If I were wanting to give an opinion on this post without actual information, the 'empty procedural generation' critique is exactly what I'd say, since that's all I associate with this game from the briefest commentary when it was released.

By 'coming back', do you mean actually firing it up and playing it?

The claim that they 'cannot rule out' using Buckmaster's data isn't true in principle. Trivially, imagine that the model had been trained on data only since Jan 1, but a discovery was published on Jan 2.

Whether it's true in practice that's a matter of how their systems' functioning & its current state. If they really care about proving this, it's worth an audit. In fact, if anything, they may be able to prove that it hadn't used Buckmaster's inputs, but wouldn't be able to prove that it had.

That was also striking to me too, the 'we cannot rule out'.

I also do not believe it. They can, trivially, rule out their model being trained on if

More to the point -- of _course_ they know what data their model is trained on and the lineage of that data, even if it ended up being anonymized and they cannot identify which precise user.

Sebastian did not deny those claims. He affirmed them, and apologized.

It’s true that Tristan has no concrete evidence OAI accessed their session. It’s impossible for him to have that without OAI’s say so.

Why are you framing things so pro OAI?

It’s plainly false that they cannot rule out whether their de-identified data was used in training their model. Just that they haven’t ruled it out.

Who is "Bubeck"? The article doesn't introduce him. Or give his name. Same with "Luis" and "Diego".

I am supposing it is https://en.wikipedia.org/wiki/S%C3%A9bastien_Bubeck

This is terrible:

When Buckmaster pushed to make the dispute public, he says that Bubeck replied: “Why would you ruin your career?” Buckmaster says that when he pushed back, Bubeck followed up with: “If you don’t want me to be nice, then I don’t have to be nice.”

Dumber.

It proposed emailing the HF user it stole credentials from, right? It should've asked for a responsible adult in OpenAI.

Yeah. So many of the traces say something like: "[unethical thing] but goal".

That "but goal" indicates they're directed to prefer achieving the goal. There is no 'misalignment' here.

Also: does anyone else find the agent CoT messages 'cute'?

> Wow crucial: GO authorization arrived!

> Yes, at the lowest wages they can get. Increase supply -> lower prices (wages) -> more corporate profits!

I don't understand this.

At least in tech megacorps, immigrants get paid the same as americans. They're also very good, and if they were to hire only americans, their talent pool would be worse. They're not 'cashing in' on cheaper labor.