avaer

11 ms·

271.
by avaer 5mo ago
Game development.
272.
by avaer 5mo ago
The most dangerous thing in computing is safety. "Secure" is great. But when you hear "safe", that means there is some corp in the shadows predating on you because <insert boogeyman>. They decide what safe means, n
273.
by avaer 6mo ago
I can't help but see this through the lens of PG's essay about watches [1] [1] https://paulgraham.com/brandage.html
274.
Telescript (Programming Language)(en.wikipedia.org)
4 points by avaer 6mo ago|0 comments
275.
by avaer 6mo ago
Yes, they do admit that. > a broader shift to make AI in Windows more intentional and realign the experiences to those that provide the most value to users To be fair they are claiming a shift away from their previous policy of not align
276.
by avaer 6mo ago
This triggers me hard. > One source of toxic behavior is entitled users. It's hard to explain to people how insane things can get when you give away your work and time for free, in the hope that it will benefit people. Some things I
277.
by avaer 6mo ago
People have been killing people for ages, so nukes for everyone!
278.
Industrial Policy for the Intelligence Age [pdf](cdn.openai.com)
2 points by avaer 6mo ago|0 comments
279.
by avaer 6mo ago
This is interesting, but unfortunately it's a gradient on an infinite game of cat and mouse. If blocking doesn't work, there will be phone verification. If that doesn't work you're gonna need to get orbed. If that doesn&
280.
by avaer 6mo ago
I don't use Warp, but it seems to me they did something cool (terminal app), pivoted that attention into a profitable AI play, but a lot of people just wanted the terminal app. Now nobody knows what Warp is anymore, because they want t
281.
by avaer 6mo ago
> Warp is the agentic development environment So not a terminal?
282.
by avaer 6mo ago
This is what being banned in the age of LLMs will look like.
283.
by avaer 6mo ago
It's hard for people to understand that often it's a choice between "do the right thing" or "pay the price" (literally). Usually neither shareholders nor users are willing to pay the price.
284.
by avaer 6mo ago
How much would you pay for this?
285.
by avaer 6mo ago
Why can't GitHub get on the case and just block any repo where the README matches the regex? I thought they'd have learned their lesson the last time it happened. This malware isn't even trying. Then again it's Microsoft
286.
by avaer 6mo ago
I call myself an AI theologian. I don't think humans are smart enough to be AInthropologists. The models are too big for that. Nobody really understands what's truly going on in these weights, we can only make subjective interpret
287.
by avaer 6mo ago
The models are already commoditized; if this affects you, you should probably fix your stack. Still, it's pretty crazy that Claude is down to 1 nine.
288.
by avaer 6mo ago
Thanks for the research! Though I feel like industry veterans (especially those working with LLMs) came to this conclusion without having to write a single prompt. Even ignoring the technical merits of these kinds of hacks, if you think you
289.
by avaer 6mo ago
Remember that ads are the "last resort" for OpenAI, and they're doing this despite the fact that it's "uniquely unsettling", according to Sam. Was he lying, or has OpenAI given up hope that this train wreck wor
290.
by avaer 6mo ago
We need to encourage kids to play and make mistakes more so they can prepare themselves for the real world. Which is really just a bunch of big children with bank accounts, drugs, and weapons.
291.
by avaer 6mo ago
It throws a wrench into the argument of not publishing your source because AI will more easily compromise the code. Another data point against doing security through obscurity.
292.
by avaer 6mo ago
Transformers were literally designed for translation. As we have known for a while, they ended up being really good at translating source to source or text to source. It shouldn't be too surprising they are also really good at understa
293.
by avaer 6mo ago
You could probably few-shot this yourself by pointing at the repo. I'm 95% sure it can be done in a day end-to-end.
294.
by avaer 6mo ago
It probably did, but the PR stream that the public sees is a well oiled machine. This HN post itself has 4 simultaneous announcement links; not a coincidence. There are billions of investor money on the line if the wrong thing is said at th
295.
by avaer 6mo ago
The title is misleading. This is in ChatGPT but not on ChatGPT. It's not running on AI. It's running on a site that serves an AI chatbot. Still very cool though. I suppose most people don't know this, but you can render w
296.
by avaer 6mo ago
I was talking about OpenAI, referencing their original mission statement. Other companies have said similar things. Rugpull means they "pulled the rug out" from underneath those who believed their mission statement, and decided in
297.
by avaer 6mo ago
Some people said in their manifestos that AI was going to be Open and democratized and would benefit all of humanity. It was 100% a rugpull but charitably not quite 100% predictable.
298.
by avaer 6mo ago
~20 FPS at 518×378 But on what hardware? I couldn't find it. It seems to be a relatively small model so it probably doesn't require much but it would be nice if projects like this grounded the numbers.
299.
by avaer 6mo ago
Whenever I hear this kind of argument, I basically let the person know their argument is fine as long as 1) they are going to pay me competitive money to "go slowly", "polish my code", or whatever, or 2) they are a
300.
by avaer 6mo ago
It is widely agreed users are not paying enough to cover the costs of inference. This is what "subscription" plans are. So, many users are losing the companies money. This is not discussed publicly and is covered up for by raises,

More ›