Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
deaton
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
91.
▲
by
deaton
3mo ago
We might as well be committing prompts given how little review happens to much of this code.
92.
▲
by
deaton
3mo ago
I think the real disaster is that once you let the LLM work on a project for a bit, you start to lose understanding of what exactly is even happening under the hood in the project. You can take steps to mitigate this, but agents don't
93.
▲
by
deaton
3mo ago
I remember when OnePlus marketed themselves as the sort of inexpensive flagship. These past few cycles they've certainly been flagship phones but I don't think there's really been a differentiator.
94.
▲
by
deaton
3mo ago
The problem is a lot of passive investors own large quantities of that ETF, and to take their money out now they have to pay a tax penalty, so they are forced to invest in SPCX due to the rule change. Its also a matter of principle. They ha
95.
▲
by
deaton
3mo ago
The one I've noticed a ton recently with Sonnet 5 is that it loves the phrase "different not in degree, but in kind." It drags that one out constantly now, at least once a day. Gemini and GPT don't at all.
96.
▲
by
deaton
3mo ago
This wouldn't even help anyway. Flock sells to law enforcement, sure, but they also sell data to everyone else who wants to know everything about everyone.
97.
▲
by
deaton
3mo ago
> You can't unilaterally say all AI blogs are low quality. That's bullshit. I can and I will say that, if not only for the fact that, in the eyes of many people, AI and the botched launch of the past few years has incurred a gr
98.
▲
by
deaton
3mo ago
I think you're right, and I think this principle of friction-is-good-actually applies to a lot more domains than just software, but whether the world will ever accept that is a different question.
99.
▲
by
deaton
3mo ago
I've always viewed unsafe Rust as a sort of last resort you use when you just can't do something safely. Porting something to "unsafe Rust" to me feels pointless, and of course this hasty rewrite was probably not the bes
100.
▲
by
deaton
3mo ago
Probably the remote work boom, in part
101.
▲
by
deaton
3mo ago
I flew Delta about 6 months ago and they had something similar, also for free, but they use Viasat. I think most of the big airlines were moving this way anyway to be honest, Starlink just has a good opportunity for advertising.
102.
▲
by
deaton
3mo ago
I live in what is probably the first place to get these things in the world, but it feels like fiber is being built at an extremely rapid pace. Just in the past couple of years it seems like Google and AT&T fiber went from being a relat
103.
▲
by
deaton
3mo ago
If AI replaces labor, there will be no money to make back
104.
▲
by
deaton
3mo ago
No it should not be, but not because of the dextromethorphan or the phenylephrine being ineffective. By far the biggest issue is the acetaminophen it contains, which it isn't super obvious about, and frequently leads to acetaminophen o
105.
▲
by
deaton
3mo ago
Torx is also common and also way way better than philips. Really we as a society need to phase out philips screws yesterday.
106.
▲
by
deaton
3mo ago
Of course there are paths to success, and they are available, but we live in a tiktok world where everyone wants to be a millionaire tomorrow and not in 20 years.
107.
▲
by
deaton
3mo ago
Its just another symptom of the silicon valley startup "do something illegal, scale fast, and dare the government to do something about it" business model. It is an absolute cancer to society, if there is such a thing as late stag
108.
▲
by
deaton
3mo ago
I cant wait for the feds to sue them for.. something
109.
▲
by
deaton
3mo ago
I think thats kinda the point. Tracking everything everyone does in an identifiable manner is extremely lucrative.
110.
▲
by
deaton
4mo ago
It probably is more secure, to be honest. I trust Google to keep my account secure more than I trust some random website to store password hashes and verify securely.
111.
▲
by
deaton
4mo ago
This is really cool, just as I'm starting to get towards the back end of the Kaishi 1.5k deck so this will be perfect for my Japanese studies. Thanks for sharing.
112.
▲
by
deaton
4mo ago
Open-weight models can be abliterated automatically with open source tools though and completely decensored. You can't do that with a closed cloud model.
113.
▲
by
deaton
4mo ago
Not really even in the same ballpark as what they did. These other labs are using AI generated content (which has already been ruled un-copyrightable) to train their models. Oh and they are paying for those tokens. So at absolute worst, the
114.
▲
by
deaton
4mo ago
Well Zai's GLM 5.2 legitimately is a frontier-level model, though not quite parallel with Opus or Fable. Unfortunately, its too damn big to run locally for most people. Thats the bottleneck right now; the open-weight models exist but s
115.
▲
by
deaton
4mo ago
With atrophy to our not-AI ability to do things
116.
▲
by
deaton
4mo ago
Really cool writeup. I personally like the Cure Dolly approach to conjugation, in which she said there was no such thing as conjugation (at least not in the Indo-European sense), only helper verbs and adjectives. Japanese grammar is just so
117.
▲
by
deaton
4mo ago
It almost seems like the juice might not be worth the squeeze. If you want verifiable code that conforms well to a well-designed plan, you have to basically write pseudocode and have the AI translate it for you. At that point why use the AI
118.
▲
by
deaton
4mo ago
It demands more and yet it makes it so, so much easier to get by with less. I don't quite understand who is supposed to be the winner here.
119.
▲
by
deaton
4mo ago
We've also seen ample evidence that AI labs are not overly concerned with the legality of how they obtain training data. Its not a stretch to say maybe they look at some other stuff they shouldn't too.
120.
▲
by
deaton
4mo ago
127,000,000 tokens used I'm sorry, I can't answer that.
More ›