Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
amelius
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
271.
▲
by
amelius
1mo ago
If this is how people in finance think, then this only makes my point stronger. The financial world is disconnected from reality and needs to be restrained.
272.
▲
by
amelius
1mo ago
Perhaps, but how is it helping the discussion by framing it like that?
273.
▲
by
amelius
1mo ago
What makes your comment better or less "populist" than mine? What does populism even have to do with it?
274.
▲
by
amelius
1mo ago
To a very limited extent, maybe.
275.
▲
by
amelius
1mo ago
You are saying it as if there is some logical consequence.
276.
▲
by
amelius
1mo ago
Doesn't matter. Banning short term trading will also reduce # trade events, so the fees don't matter.
277.
▲
by
amelius
1mo ago
For one thing because now people who own high-speed low-latency connections to stock markets have no advantage over "normal" people who actually do meaningful work, like being a policeman or nurse.
278.
▲
by
amelius
1mo ago
If this whole efficiency argument means I'm going to make $1 more per month while some lousy traders make $100k more per month, then I'd say: to hell with "efficiency". Fairness and wealth equality are more important to
279.
▲
by
amelius
1mo ago
How much will it hurt me? Maybe I'm OK with being hurt through the markets if it means the world is a saner and more fair place?
280.
▲
by
amelius
1mo ago
Well, long term trading is what we should stimulate because it corresponds to investments based on insights in the underlying companies. Short term trading, however ... Maybe we should just put limits on how short one can own a stock (?)
281.
▲
by
amelius
1mo ago
Now try to do them in safe Rust.
282.
▲
by
amelius
1mo ago
Since at least Windows 3.1 we all know that cooperative multitasking is a bad idea. But we see a lot of developers use it, even Rust developers who should know better.
283.
▲
by
amelius
1mo ago
(I'm not saying there couldn't be improvements in the area of the actuators)
284.
▲
by
amelius
1mo ago
I don't think that's true. AI could work around all the physical imperfections. Just imagine yourself controlling an imperfect robot with some joysticks or maybe a sensor suit. You'd certainly do much better than current robo
285.
▲
by
amelius
1mo ago
Can we have a unicode character that means "don't use for AI training", and another one for "this is AI generated data", and while we're at it one for "don't use for advertising purposes".
286.
▲
by
amelius
1mo ago
It's not in big-money's interest to call them conscious because that would mean they get rights too.
287.
▲
by
amelius
1mo ago
But it seems to me that the whole failure could have been prevented with a bunch of temperature sensors spread throughout the engine, and a good anomaly detection system.
288.
▲
by
amelius
1mo ago
It looks nice but it also saddens me a bit. You did all the easy and nice work to design a nice form factor and make a nice website, but you left all the hard work for the customer ... Not sure what to think. Maybe if the marketing was more
289.
▲
by
amelius
1mo ago
What if you use the model to write children's stories and it starts regurgitating Mein Kampf? Having some knowledge about the training data is useful.
290.
▲
by
amelius
1mo ago
If you're worried about safety: compliance can be implemented outside of the motor, e.g. with springs.
291.
▲
by
amelius
1mo ago
What I want is a model that is trained with data that is openly available, where the data is curated by academia. I don't want corporate crap in my AI (unless it has been filtered properly).
292.
▲
AI Slop Is Ruining Cute Animals on the Internet
(wired.com)
8 points
by
amelius
1mo ago
|
0 comments
293.
▲
by
amelius
2mo ago
Well, if it can fold my laundry and do the dishes, then jerky movements don't matter to me. I suppose precision can be fixed with camera feedback. And those force feedback actuators use significantly more power.
294.
▲
by
amelius
2mo ago
What does your boss say? Do they think the company is doing fine?
295.
▲
by
amelius
2mo ago
How do you know for sure that you stay on an efficient frontier when you change a parameter? I think this presentation says more about what knobs you can turn and in what direction the outcome will move (it may be worse than a competitor) t
296.
▲
by
amelius
2mo ago
Where can you buy them?
297.
▲
by
amelius
2mo ago
> The main reason to run local: cloud APIs are rented land. They can change their pricing, hit your usage limits, or swap the model being served behind the scenes whenever they feel like it. Yes but it's easy to replace them. The ma
298.
▲
by
amelius
2mo ago
It's a pity they use the nabla symbol for vector difference. This symbol usually has a different meaning in the context of vector calculus.
299.
▲
by
amelius
2mo ago
200 MHz is good enough for most things. And for modern computing electronics even 1GHz is too slow, e.g. monitoring USB3 or high speed PCIE data lanes. Did you get a price quote for that keysight oscilloscope? I see that you can also rent t
300.
▲
by
amelius
2mo ago
How usable is 12 tok/s?
More ›