Bragging rights and improved sleeping ability from the knowledge that the devs are being supported.

The serious answer is it’s often easier for people in a company to buy a license key than it is for them to arrange a donation to the devs. So this is an easy way to make small donations.

Well it is trained to copy musk.

I guess it might work if HR don’t know how an LLM works. There’s not many that can edit a word file so it includes whited-out footnotes.

You’re better off getting a friend to lie for you. They can say they added it while helping you with formatting and you know nothing about it.

Genuinely, this already happens in large companies for related reasons.

The CV is on file, and if HR reprocess it for any reason e.g. relocation or change of role, it’s automatic dismissal for dishonestly if they catch a deliberate lie.

No but they can fire you later even if you’re good at the job.

Then you’re stuck in an even worse position with a big gap in your CV and no reference.

Unfortunately, this is seen as dishonestly and is grounds for immediate dismissal in a lot of places.

In practice it’s very systematic for small networks. You perform a search over a range of values until you find what works. We know the optimisation gets harder the deeper a network is so you probably won’t go over 3 hidden layers on tabular data (although if you really care about performance on tabular data you would use something that wasn’t a neural network).

But yes, fundamentally, it’s arbitrary. For each dataset a different architecture might work better, and no one has a good strategy for picking it.

Probably because there’s no good reason.

At least one intermediate layer is needed to make it expressive enough to fit any data, but if you make it wide enough (increasing the blobs) you don’t need more layers.

At that point you then start tuning it /adjusting the number of layers and how wide they are until it works well on data it’s not seen before.

At the end, you’re just like “huh I guess two hidden layers with a width of 6 was enough.”

It’s interesting. There’s a lot of talk about how chatgpt makes people lazy, but honestly I think Google killed the “read the manual” ethos.

Back in the day when you couldn’t just search for everything, you needed enough understanding of the manual to find anything in the index.

So a key part of figuring anything out was reading at least the start of the manual.

Now, fuck it, you just type into Google and try to guess enough context to understand what’s going on.

Yeah, but it can be filled in with “house”.

Hexbear is already flooded with beanis posts.

Looking forward to seeing beanis everywhere in the next version of Facebook’s LLM.

Garmin sends all your health data to the cloud and the app won’t work without an Internet connection.

On the plus side, they’re not part of the Google/Apple/Samsung data ecosystems, and I don’t think actually they do anything with the data, beyond computing statistics for you.

Depends how much you’re prepared to trust them I guess.

You’ve got it backwards.

This kind of voting forces the existence of two party systems.

Suppose you have two parties one left wing that gets 60% of the vote and one right wing that gets 40% in every district. Right now the left wing always wins.

If the left wing party splits into two blocks of 30% each, the right wing always wins.

So if you want to win, you can never split from the large parties, even if you feel unrepresented.

Depends.

You can argue that it’s basically art/political speech. You’ve done it to draw attention to flaws in the approach and to highlight how ineffectual the current system is, and that if you actually wanted to do make fake IDs you’d take a much less high-profile approach. As such, there’s no actual criminal intent required.

Don’t know if a judge would buy it though.

Selfie doesn’t work, you need to turn your head left and right to follow instructions.

But yeah, there’s a bunch of avatars that will bypass it

Ok, but one of the most important use cases is non-local access.

If I’m at home I can just go to the door.

Go on, drop a rocket on Zuck’s Bond villain hideout.

Let’s see what happens.

Tesla and to a lesser degree spaceX are where the money comes from.

Everything else, Twitter, xAi, the boring company are all vanity projects.

We’re seeing the start of the tipping point. Tesla is being substantially harmed by his personal fuckery, but it’s always been a meme stock, so the price is much higher than it should be. This means that when the crash finally comes it should be extremely messy.

SpaceX is already being used to bail out xAi - if Tesla was healthy musk would rather use it’s funds instead - and with musk fighting with Trump they could get some contracts wiped out. They’ve got enough actual products (star link) without competitors that they’re probably safe for now.

Commercial versions of these systems exist in the UK.

https://www.theguardian.com/uk-news/2025/jun/06/shopper-facewatch-watchlist-39p-paracetamol-london

The Gdpr makes these things harder to do, but not automatically illegal.

Surely you have noticed that there is a lot of criticism of the GDPR and EU tech regulation.

Yeah, and some of it is even true.

So a random person on Reddit claimed there’s about 800 million possible uk mobile numbers, some people have multiple numbers so ballpark 80 million active phone numbers. This gives around a 1:10 chance of picking an active number at random. If there’s actual patterns in the numbers this could be even more likely.

What’s interesting is this won’t have been a realistic sounding number.

Company lines typically start 0300 or 0800 but mobiles are 07… Something.

So if it was just hallucinating, it did so badly.