They will compensate agent engineers just like they compensated all of the developers, artists, academics, and editors that created the data models are trained on.

> let's do this for politicians as well :), publicly available, auditable 24-hour surveillance

Politicians will be the first to carve out exceptions for themselves for reasons of "security" while everyone else is surveilled.

Yes, it should literally be the opposite -- with power should come accountability. But that's not how these things work in practice.

If we are living in a democracy and in majority are able to live democracy, then yes. If on the other hand most people are not willing or unable to live democracy, then we will get stuck with corruption.

How do you remove them?

This data is going to get leaked in a breach. It will be used against you in a court of law. It will be used for training and (regardless of what anyone says) will be used to fire you once the AI can do your job.

And when all of the above happens Meta will be absolved of any responsibility.

I don't understand how it's legal either. I guess we need laws against it yesterday.

It doesn't have to get leaked. They can sell it and use it as another means to identify Internet users. Meta is pretty infamous for identifying, tracking, and understanding user behavior. We are kind of past the point where these companies care at all. If you think the push to add age verification to operating systems is an unrelated giggle I envy you. Something something Cambridge analytica.

I think it's their employees here that have cause to be concerned, not internet users.

Meta already has literally have billions of people's personal profiles and browsing history.

I don't think screenshots of their SWE's IDEs is going to be useful for identifying internet users.

They could perfect it in house and then roll it out as a product. The way people type and use a mouse are pretty identifying especially when coupled with other things.

I do agree screenshots themselves are less useful for that.

That doesn't make any sense.

1. Why use their employee's data to fingerprint input? They could do that to a billion+ of their users instead.

2. Input fingerprinting is multi-decades old science, there are already production products that do this.

Are there products that do this with all of the other metadata that meta now collects? At the scale that meta collects them? My guess is no.

I would be highly surprised if they don't do this already for bot detection... but again, if they want to do it to track people on the internet, the data that would be useful is data from the internet, which they have an incredible wealth of -- not a dataset that is several orders of magnitude smaller from their internal employees' desktops.

Same with employee surveys, BTW.

Never trust that you are anonymous.

That sounds actionable if your lawyer (that you're paying) isn't actually working for you.

Wait until you talk to the state Bar. That is when the real fuckery starts.

Good luck getting a lawyer to sue another lawyer either.

You could judge the costs of the AI products you're using by the standard API pricing, not promotional subscription offers.

Not even that way, given that the price is still highly subsidized by investors and circular deals.

For me, it's not even cost necessarily. If they decide to change the product they offer, the old one is gone. I refuse to use anything for personal use that's not at least _available_ as model weights.

Whether art is considered good is in practice highly contextual. One of those contexts is who (what) made it.

Well, that's the entire point.

If we take a completely utilitarian and amoral viewpoint, the insiders are selling their material, non-public information. The rest of the market participants are buying. From the latter's utilitarian perspective the former are providing a valuable service and getting paid for it. I'm pretty sure there was a legal term for such sales activity...

The only people playing fair are those who don't know how to cheat well enough.

If it's a PR stunt it's a good one. I hadn't heard about mulerun before this.

The article casts doubt on the claim of no experience.

He references actual research papers, complains about regressions, and attempts to shape swarms.

You decide for yourself.

> You'll get a competent UI with little effort but nothing truly unique or mind-blowing.

The shelf-life of unique and mindblowing has reduced to a week (being generous) before it's copied by slop artists looking for a resume booster or funding, and months tops before it's part of training data for everyone. Unless you find it in that small time window everything will seem homogenous.

It could just be a systemic result; unless you deliberately take the lonely road to parts of the internet where other people aren't, you will not see unique and mind blowing things. Which by definition you can't source from a place that has a lot of users, like social media or popular forums.