joegibbs

9 ms·

181.
by joegibbs 1y ago
Nettles have a lot of uses. Back in the day my grandfather's neighbour would, whenever he was getting sick, take off all his clothes and roll around in the nettles. I've got no idea why - he kept doing it so it must've done s
182.
by joegibbs 1y ago
How do the models do if you ask them to translate to X language and adapt the text to suit cultural norms and idioms?
183.
by joegibbs 1y ago
They occasionally do stupid things like this, the other day I asked Codex to make some changes to a few files and it refused because it was too much work.
184.
by joegibbs 1y ago
It’s a fundamentally unsolvable philosophical problem since it relies on knowing about things outside the universe, which is impossible. Say the outside universe had entirely different laws of physics, you would have no idea. It’s also basi
185.
by joegibbs 1y ago
I notice this less with GPT-5 and GPT-5-Codex but it has a new problem: it'll write a sentence that mostly makes sense but have one or two strange word choices that nobody would use in that situation. It tends to use a lot of very dens
186.
by joegibbs 1y ago
How does that help though, if a charity uses an entirely raw unedited photo that is still totally fabricated or misleading? I think that’s the real problem - it’s not really any better if a charity uses a paid actor on a set, or finds the a
187.
by joegibbs 1y ago
For one, like the article says, there’s the fact that it’s not really news: you could print "BREAKING: 5,000 nonagenarians DEAD from heart disease and cancer" every day since people started living that long. It’s a shock when a yo
188.
by joegibbs 1y ago
If the user needs to set their own width by resizing the window they can also set the width by right-clicking -> inspect element -> disabling the CSS style on the div that gives it a max width. Which is only a couple more clicks than
189.
by joegibbs 1y ago
You'll also notice the same thing happens for other non-existent emojis that sound like they should exist: dragonflies, lemurs, possums, blackberries - even Claude 4.5 will start off by saying "Yes!" and then correcting itsel
190.
by joegibbs 1y ago
God I can't stand it when I get this kind of output from Claude, they really need to train it out for Claude 5. "[Tangentially related emoji] I have completed this fully functional addition to the project that is now working perfe
191.
by joegibbs 1y ago
Not even them, simulations of them . They're not even real! Why the hell should I care if some far-future AI tortures simulations of me? Go ahead, spin up a hundred, hell spin up a trillion!
192.
by joegibbs 1y ago
And it wasn't as good as it's often mythologised to be. Back in the 60s (in Australia) people weren't going on holidays overseas, they lived in houses under half the size of modern builds, they had worse healthcare and lived
193.
by joegibbs 1y ago
It depends on the salary, right? If you're in Silicon Valley paying 500k TC it probably makes sense to let your employees go wild and use as much token spend as they like.
194.
by joegibbs 1y ago
I remember that a couple of years ago people were talking about how multimodal models would have skills bleed-over, so one that's trained on the same amount of text + a ton of video/image data would perform better on text response
195.
by joegibbs 1y ago
There are a lot of sci-fi inflected ideas parroted in the media that are the cause of this. For dramatic purposes the AI in sci-fi has to act as an autonomous living being just like a human with its own goals and desires. This doesn’t map a
196.
by joegibbs 1y ago
But what exactly does this product do that you can’t from just parsing the stream? Besides, the problem with hallucinations is the unknown unknowns: if what you’re doing is easily verifiable (like parsing JSON or checking valid chess moves)
197.
by joegibbs 1y ago
Same here. I'd never seen parts of it on TV or anything until I watched it for the first time, since I'd seen it was ranked #1 on IMDB. There wasn't anything about it that made me feel like it was the best movie ever . The p
198.
by joegibbs 1y ago
We hear a lot about things that decrease blood flow to the brain, but are there any drugs that increase it? And would there be any benefits for people who don’t already have decreased blood flow, or does the brain naturally just get as much
199.
by joegibbs 1y ago
I don't think this will work. There are too many tools, a lot of them legacy, and it will be an incredible pain to get them all coordinated on a new standard. Instead I think it'd be better to solve this on the agent side by check
200.
by joegibbs 1y ago
Where's the link? Why is everyone upvoting this when there's no story attached?
201.
by joegibbs 1y ago
Short term pain, long term gain. People should absolutely not be making friends with a software program. Don’t get me wrong, I’m not even anti AI. But you should absolutely not be replacing your human connections with a statistical model th
202.
by joegibbs 1y ago
In business terms it makes perfect sense - if people care more about the brand and image than the actual reliability and quality of the product, why bother spending so much money on quality? Of course the brand will end up losing its appeal
203.
by joegibbs 1y ago
It wasn't the entire heart but people have had things like valves transplanted from animals. Kevin Rudd had a bovine aortic valve transplanted back in 2011.
204.
by joegibbs 1y ago
Cool! I'm working on something similar ( https://app.embedsheet.ai , a bit buggy and in-dev, the paid stuff is just placeholders) but yours feels much more responsive and smooth than mine, it's very impressive! I notice
205.
by joegibbs 1y ago
Has there really ever been such a thing as a benevolent country? I think it only exists in subjective histories depending on who wins. For instance the Romans are thought of as a great, cultured civilisation because their histories won out,
206.
by joegibbs 1y ago
If it's a decentralised non-profit platform which doesn't have future plans to monetise the users somehow, isn't spending $0 on marketing the default? You'd hardly want to throw a few mil into ads for something that will
207.
by joegibbs 1y ago
I disagree on the AI 2027 thing, I don't think that we've seen any evidence of self-improvement or a rate of increase in abilities that it suggests. If anything I think that the rate of improvement has slowed since GPT3, without s
208.
by joegibbs 1y ago
https://archive.is/20250822002317/https://www.theatlantic.co...
209.
The AI Doomers Are Getting Doomier(theatlantic.com)
4 points by joegibbs 1y ago|3 comments
210.
by joegibbs 1y ago
If the LLM is multimodal would more video and images improve the quality of the textual output? There’s a ton of that and it’s always easy to get more.

More ›