xpct

Read through most of this. It does sound like fun detective work, looking at what neurons fire. Some of the concepts get quite complex. I wonder if this can be gamified.

My favorite thing about puzzle games is meta, cognition stuff. What mindset can I put myself in, so that I can solve this faster? What should I think about, and how, to reduce complexity of the problem?

I tried the same last summer with Baba Is You and had an absolute blast. Playing through OotSS now, and momentum has been solid for the past ~20 hours. Skipping Stones has been my favorite!

Indeed. I'd like to remind everyone that with LLMs it's increasingly simple to build quite a good profile of someone just from their social media posts.

I wonder what the downstream effects of that will be. I imagine we may see some big lawsuits soon, meant to scare off the RE trend.

I don't think this is a statement about competition. This is not a metric anyone has competed on so far, it's being made into a press release.

> It seems that the solving rather than just the solution gave much to mathematics

The whole discourse of past few weeks has been mathematicians claiming the opposite.

The 'vibe coded' part needs to go. Both teams likely used AI-assisted coding. Neither team's product was completed by robots, not with current tech. Epic had the launcher update ready in advance to handle the bad PR from their announcement. Prepare for more announcements from Epic how AI and not humans solved something.

I'd hope it doesn't get solved, because it's one of the few signals I have to the person behind the screen.

I don't think you should feel entirely safe in warring nations, no matter what you do.

Hmm.. I meant to make my point in isolation from LLMs. I definitely don't feel that way about learning music.

I do kind of understand it when it comes to programming, though. Hobby-like side projects often felt like they had some utility down the line, even if not directly applied. It's fun to show others what you've accomplished, to show off, maybe land a job from being noticed, etc.

I don't know if it's common but I generally don't like to be bad at my hobbies. I like it when I can do things well. I've been learning the piano recently and striving to play more difficult songs is what makes it fun.

Happy to hear that. What are you speedrunning?

Where fruit flies carry cameras around? That's scary.

Likely at some point regular people just aren't given access to such edge tech, though. Even with infinite knowledge from an LLM, you probably couldn't build it without access to a factory either.

Well otherwise you are hiring people who are ill-intentioned to make others addicted to dogfood.

Even then, just put a fake bottom in and nobody would notice. Worst case they throw out your mic.

007 -> bondslop

I'm quite sure some won't bother.

I think the line is wherever you feel comfortable with pushing the changes yourself. You feel like adequate testing had been done, basing it on previous testing periods or how successful AI runs have been.

Whether you or AI wrote the code, you decide when it's safe enough to put your name on it. Like it's always been!

I'd love to see an estimate of how much Software Engineering results in jobs being automated. I'd bet it's not that big of a number, and part of that reason that classical coding approaches couldn't close most loops end to end.

That by itself doesn't imply need for infinite code. It may just imply the opposite.

Could you point me to something that their harness does exceptionally well as compared to their competitors, and why that can't be accomplished by a third party?

I don't think anything points to Anthropic having the best software engineers, at least their products surely don't!

Correct me if I'm wrong, but I don't think this has ever been the dream? The dream was to write software once, and then use it forever.

Frankly, I don't think the main concern is children. A lot of adults don't have good impulse control, and we should build systems that disincentivize harmful behavior. I'm more worried about someone doomscrolling fetish subreddits today than having to contact sketchy businesses back in the day, because of the hurdle they had to get over.

I think consuming sex content is akin to most time-wasting we can do on the internet today, and that'd be my main concern.

That's been my thoughts as well, but I'm sure at least some of it can be repurposed.

Yes, the friction has been a lot lower, and that builds different habits. You can reach a lot of stuff with the click of a button, sometimes even accidentally on social media such as X or Reddit. Given that we have video tech now, that's also likely much more alluring than images.

Search engines were very interesting and had a lot of myths surrounding them. People claiming that typing out full sentences worked better, vs others entering just keywords. A lot of people learned to utilize the special operators.

I thought I got really good at googling right before the engine started becoming really bad around Covid years.

I'd love to read something about the topic.

I don't necessarily support banning it, but I don't think it's been all positives from how easy to access adult content has become over the years.

It was my impression that some of these companies are getting acquired for the data they collected, and $20M seems on the low-end nowadays, especially given the traffic they had.

So, is it that Disney isn't getting the data, is the data poor, or is it just not a data thing at all?

If it uses different methodology I'd be happy to see it on HN, even if it doesn't include the flavor of the month model.