Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
I_am_tiberius
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
91.
▲
by
I_am_tiberius
8mo ago
I once tried hugging face because I wanted I worked through some tutorial. They wanted my credit card details during the registration as far as I remember. After a month they invoiced me some amount of money and I had no idea what it was. T
92.
▲
by
I_am_tiberius
8mo ago
I hope it's Fairphone.
93.
▲
by
I_am_tiberius
8mo ago
He's an experienced founder (psPdfKit)!
94.
▲
by
I_am_tiberius
8mo ago
I really hoped he would support Europe’s startup ecosystem. Hopefully, he will at least bring stronger privacy standards to OpenAI, such as a policy that prohibits reading or analyzing user prompts or AI responses.
95.
▲
by
I_am_tiberius
8mo ago
If it's in your power, make sure user prompts and llm responses are never read, never analyzed and never used for training - not anonymized, not derived, not at all.
96.
▲
by
I_am_tiberius
8mo ago
I think the only danger is quantum computers.
97.
▲
by
I_am_tiberius
8mo ago
I'd like to know if and how much illegal use of customer prompts are used for training.
98.
▲
by
I_am_tiberius
9mo ago
It's certainly unethical to have used the naming in order to get on the hype train. This was clearly a strategic decision.
99.
▲
by
I_am_tiberius
9mo ago
My assumption is that it goes the other direction on a permanent basis.
100.
▲
by
I_am_tiberius
9mo ago
> "WhatsApp provides default end-to-end encryption for over 3 billion people". Wasn't there news lately that they can still read your messages somehow?
101.
▲
by
I_am_tiberius
9mo ago
Yes I know! Scaleway is great as well. But I was referring to the product portfolio.
102.
▲
by
I_am_tiberius
9mo ago
I agree with Scaleway (I would more compare it to Digital Ocean) but OVH is really good and comparable.
103.
▲
by
I_am_tiberius
9mo ago
Pity it doesn't support other llms.
104.
▲
by
I_am_tiberius
9mo ago
I don't get the pricing page. It seems not to have prices.
105.
▲
by
I_am_tiberius
9mo ago
+1
106.
▲
by
I_am_tiberius
9mo ago
That thing is too big for me. Also, I want the mousepad in the middle of the device and not left aligned.
107.
▲
by
I_am_tiberius
10mo ago
I want a linux laptop with more than 64gb of ram.
108.
▲
by
I_am_tiberius
10mo ago
Aren't you watching the news? Trump himself says it.
109.
▲
by
I_am_tiberius
10mo ago
The only interest Americans have is oil. nothing else.
110.
▲
by
I_am_tiberius
10mo ago
Because we want the best for other people.
111.
▲
by
I_am_tiberius
10mo ago
- Nix OS - French
112.
▲
by
I_am_tiberius
10mo ago
I'd have expected Fairphone to support GrapheneOS:(.
113.
▲
by
I_am_tiberius
10mo ago
I still remember the 3 level demo CD they made. Awesome demo levels that made me buy the game, which is still one of my favorite games of all time.
114.
▲
by
I_am_tiberius
11mo ago
I said it and say it again, openai is the biggest privacy disaster ever. Sam Altman should be ashamed, because it's not at all necessary.
115.
▲
by
I_am_tiberius
11mo ago
No, when memory is disabled this shouldn't be the case.
116.
▲
by
I_am_tiberius
11mo ago
Some months back they changed their terms of service and by default users now allow Anthropic to use prompts for learning. As it's difficult to know if your prompts, or derivations of it, are part of a model, I would consider the possi
117.
▲
by
I_am_tiberius
11mo ago
Still mad at them because they decided not to take their users' privacy serious. Would be interested how the new model behaves, but just have a mental lock and can't sign up again.
118.
▲
by
I_am_tiberius
11mo ago
I guess you should never use the latest versions of libraries.
119.
▲
by
I_am_tiberius
11mo ago
I don't know if this is true but I believe Anthropic has for a long time illegally used user prompts for training, without user consent.
120.
▲
by
I_am_tiberius
11mo ago
I've never been asked a phone number. Maybe country specific. no idea.
More ›