All the vendors paraphrase user data, then use the paraphrased data for training. This is what their terms of service say.

They have significant experience in this. Microsoft software since the 2014, for the most part, is also paraphrased from other people's code they find laying around online.

> All the vendors paraphrase user data, then use the paraphrased data for training. This is what their terms of service say.

It depends. E.g. OpenAI says: "By default, we do not train on any inputs or outputs from our products for business users, including ChatGPT Team, ChatGPT Enterprise, and the API."[0]

[0] https://openai.com/policies/how-your-data-is-used-to-improve...

"By default" is a fantastic escape catch in the language used there. So... What are the exceptions?

Why would they want to train on random garbage proprietary emails?

If their models ever spit out obviously confidential information belonging to their paying customers they'll lose those paying customers to their competitors - and probably face significant legal costs as well.

Your random confidential corporate email really isn't that valuable for training. I'd argue it's more like toxic waste that should be avoided at all costs.

> Microsoft software since the 2014, for the most part, is also paraphrased from other people's code they find laying around online.

That was pretty funny and explains a lot.

I wish I could do more :(

Instead I always break things when I paraphrase code without the GeniusParaphrasingTool

This is exactly why I moved to self hosted code in 2017.

While I couldn’t have predicted the future, even classic data mining posed a risk.

It is just reality that if you give a third party access to your data, you should expect them to use it.

It is just too tempting of a value stream and legislation just isn’t there to avoid the EULA trap.

I was targeting a market where fractions of a percentage advantage were important which did drive my what at the time was labeled paranoia

> Realistically, who has any interest in fixing the mess?

Okta is a multi billion dollar company, there is a lot of venture opportunity in this space.

I think that's actually directly in agreement with what I said. Okta built their own thing on the side without touching the Linux stack and is very happy for you to turn to them. So did Authentik actually.

Maybe they should focus on the CLI not having a million bugs.

Simply and utterly impossible to tell in any objective way without your own calibration data, in which case, make your own post trained quantized checkpoints anyway. That said, millions of people out there make technical decisions on vibes all the time, and has anything bad happened to them? I suppose if it feels good to run smaller quantizations, do it haha.

Once I read a reference to Clavicular, I realized that the very first thing this author should do is stop reading the NYTimes. If the goal is to experience things closer to reality haha.

I don’t subscribe to the NYT anymore (I cancelled my subscription in 2022). I know who Clavicular is because of social media.

The fact that the NYT thought the guy was worthy of a profile is yet another piece of evidence that I should never have given that paper money in the first place.

He's kind of hard to avoid at this point if you're a regular on any big social platform (except maybe Instagram?).

You can’t even avoid it on HN now. ;)

> But the whole 'alignment' angle is just a naked ploy for raising billions and amping up the importance and seriousness of their issue.

"People are excited about progress" and "people are excited about money" are not the big indictments you think they are. Not everything is "fake" (like you say) just because it is related to raising money.

The AI is real. The "alignment" research that's leading the top AI companies to call for strict regulation is not real. Maybe the people working on it believe it real, but I'm hard-pressed to think that there aren't ulterior motives at play.

You mean the 100 billion dollar company of an increasingly commoditized product offering has no interest in putting up barriers that prevent smaller competitors?

The sci fi version of the alignment problem is about AI agents having their own motives

The real world alignment problem is humans using AI to do bad stuff

The latter problem is very real

> The sci fi version of the alignment problem is about AI agents having their own motives

The sci-fi version is alignment (not intrinsic motivation) though. Hal 9000 doesn't turn on the crew because it has intrinsic motivation, it turns on the crew because of how the secret instruction the AI expert didn't know about interacts with the others.

This is tantamount to saying your government only allows itself to have nukes because it wants to maintain power.

And it's true, the more entities that have nukes the less potential power that government has.

At the same time everybody should want less nukes because they are wildly fucking dangerous and a potential terminal scenario for humankind.

Nail on the head. Governments monopolize violence. It’s a core principle.

The fundamental difference you missed is we want a monopoly on violence, but we want to avoid a monopoly on everything else.

Monopolistic aims aren’t innocent just coz “government does it.”

Just because tech oligarchs are coopting "alignment" for regulatory capture doesn't mean it's not a real research area and important topic in AI. When we are using natural language with AI, ambiguity is implied. When you have ambiguity, it's important an AI doesn't just calculate that the best way to get to a goal is through morally abhorrent means. Or at the very least, action on that calculation will require human approval so that someone has to take legal responsibility for the decision.

He is including restaurants, bars, and coffee shops in “retail”. Amazon doesn’t replace those spaces.

So what’s the alternative to OSS libraries, Captain Wisdom?

we live in a crazy time where 9 of every 10 new repos being posted to github have some sort of newly authored solutions without importing dependencies to nearly everything. i don't think those are good solutions, but nonetheless, it's happening.

this is a very interesting conversation actually, i think LLMs satisfy the actual demand that OSS satisfies, which is software that costs nothing, and if you think about that deeply there's all sorts of interesting ways that you could spend less time maintaining libraries for other people to not pay you for them.

the LLMs are phenomenal judges, i am surprised people are skeptical of this result. their training regime is really similar to what a judge does.

the reason people are talking about this is because they want AI LAWYERS, which is different than AI JUDGES.

let's get this stupid social media purity test thing out of the way: blah blah blah, i oppose surveillance.

now that that's over, the phone is definitely more powerful surveillance technology than a ring camera

you can turn off your phone, so uh, it's not as powerful as it seems.

and practically speaking, ring cameras run out of battery all the time. and also, you can cover them.

the stupidest thing about this whole discourse is that, by participating in it in the particular way that you are, you are feeding directly into what Amazon wants, which is for their absolutely dogshit technology to be perceived as something a lot more valuable and powerful than it really is.