Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Nition
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
14 ms
·
271.
▲
by
Nition
1y ago
Interesting, my piano is a Yamaha (specifically a 1980 Yamaha YUA, basically a U3), and it does follow the model (same width everywhere). Photo: https://i.imgur.com/ulsLUoG.jpeg I also have a MIDI keyboard (M-Audio Hammer
272.
▲
by
Nition
1y ago
The way I've dealt with that in the past is putting into into Review or whatever the equivalent is, make a note ("cannot repro, but attempted potential fix in version XXXX, moving to review, please reopen if anyone reports this ag
273.
▲
by
Nition
1y ago
It's not just one report from 17 years ago, it's 194 comments with the most recent one from nine months ago. It doesn't seem like mitigation steps have been implemented.
274.
▲
by
Nition
1y ago
Thanks for linking that. I've tried Thunderbird a couple of times in the past and quite liked it, but that thread has put me off using it forever. Even if the bug is fiendishly hard to track down and reproduce, you'd think there w
275.
▲
by
Nition
1y ago
Yeah, my suspicion is that current-style LLMs, being inherently predictors of what a human would say, will eventually plateau at a relatively human level of ability to think and reason. Breadth of knowledge concretely beyond human, but inte
276.
▲
by
Nition
1y ago
I took that comment as more like "it doesn't have any effect beyond the output of the model", i.e. unlike saying something like that to a human, it doesn't actually make the model feel anything, the model won't spre
277.
▲
by
Nition
1y ago
You guys have got stuck arguing without clarity in what you're arguing about. Let me try and clear this up... The four potential scenarios: - Mild prompt only ("no orange cats") - Strong prompt only ("no orange cats or p
278.
▲
by
Nition
1y ago
Because prompts are never 100% foolproof, so if it's really life and death, just a prompt is not enough. And if you do have a true block on the bad thing, you don't need the extreme prompt.
279.
▲
by
Nition
1y ago
This is correct of course but we would only confuse the Americans.
280.
▲
by
Nition
1y ago
This looks really nice (apart from the shape of the gibbous moon, but others have already pointed that out). Could you possibly detect the user's approximate location and rotate the whole thing 180 degrees if they're in the southe
281.
▲
by
Nition
1y ago
The SomethingAwful forums have been $10 to join since 2001. If you get banned, that's another $10 needed for a new account. Has worked pretty well for 24 years.
282.
▲
by
Nition
1y ago
I have a similar thought re "work sucks"/"nobody wants to work anymore", where actually maybe work and people are the same as ever but being at home is more entertaining than it used to be.
283.
▲
by
Nition
1y ago
Yeah, but the challenge in this case was to achieve it without any rounded edges.
284.
▲
by
Nition
1y ago
My pet theory was that they wanted to stay on the next version indefinitely but Mac OS seemed to be indefinitely on version 10, and they couldn't be one behind. Funnily enough, soon after Mac OS went to version 11, Windows 11 was annou
285.
▲
by
Nition
1y ago
I have to say it made the later "Runtime Fee" announcement seem even more in poor taste as well, given that it might have had a big effect on the users paying it, but ultimately mean almost nothing to Unity against the billions pa
286.
▲
by
Nition
1y ago
The fact that they were struggling for revenue just made the massive spend seem even weirder to me, but I suppose it could make sense if they truly expected to somehow get >4.4 billion back from ad revenue eventually. They also bought Wē
287.
▲
by
Nition
1y ago
The weirdest part of that merger was Unity paid $4.4billion for IronSource.
288.
▲
by
Nition
1y ago
Yeah this is the big benefit of current LLM AI even with the mistakes and hallucinations IMO. All the things that are hard to answer but easy to verify. Not just programming. e.g. You have a complex medical problem. Hard to ask Google. Ask
289.
▲
by
Nition
1y ago
The good option would be for the LLM to say it doesn't know. It's the making up answers that's the problem.
290.
▲
by
Nition
1y ago
https://www.youtube.com/watch?v=B7MIJP90biM
291.
▲
by
Nition
1y ago
I suppose 'vibe coding' kind of is macrodata refinement is a way.
292.
▲
by
Nition
1y ago
That makes a lot of sense; thanks for the perspective from the artist's side.
293.
▲
by
Nition
1y ago
Tangentially related: Recently I was thinking of commissioning an artist to do some album art for me. I had a specific concept in mind and it needed to have a certain look to it but I didn't mind if the actual art was physically painte
294.
▲
by
Nition
1y ago
> increasing energy demands for AI: companies shoving it into products that nobody wants or needs I think of this a little every time Google gives me another result with the AI summary and no option for anyone to turn it off. Apparently
295.
▲
by
Nition
1y ago
While we're listing complaints... 250MB for a screen recorder update?
296.
▲
by
Nition
1y ago
I've noticed a current problem in Win 11 where if you start your PC with the monitor off, let it get to the login screen, then turn the monitor on, you get a terribly low-res version of your login window background. I suspect it's
297.
▲
by
Nition
1y ago
The process I'm suggesting is: 1. You have a complex or vague question that you can't search easily via Google etc 2. You ask the AI and it converts that to concrete searchable suggestions (in this case "sudo rm -rf /&qu
298.
▲
by
Nition
1y ago
Grok 3 Marathon Crater, likely a confusion with Marathon Valley on Mars, wasn't "discovered" in the traditional sense as a distinct crater. Marathon Valley is a region within the western rim of Endeavour Crater, explored
299.
▲
by
Nition
1y ago
Medical was just one example, replace with anything you like. As another example, you can give the AI a photo of something to have it name what that thing is. Then you can check the thing by its name on Google to see if it matches. Much eas
300.
▲
by
Nition
1y ago
One type of question that a 20%-failure-rate AI can still be very useful for is ones that are hard to answer but easy to verify. For example say you have a complex medical problem. It can be difficult to do a direct Internet search that cov
More ›