Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
tempaccount420
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
61.
▲
by
tempaccount420
10mo ago
It was React. Code in React's repository had to be patched to fix this. Next.JS just happens to be the biggest user of this part of React, but blaming Next.JS is weird...
62.
▲
by
tempaccount420
10mo ago
They should be performantly removed.
63.
▲
by
tempaccount420
11mo ago
It's never JavaScript itself, but somehow it's almost always JavaScript code...
64.
▲
by
tempaccount420
11mo ago
Basically, JavaScript should not be running on servers. Vulnerabilities caused by shoddy JS are a lot more impactful to a server since multiple users will be served by the same runtime instance.
65.
▲
by
tempaccount420
11mo ago
Gemini 3 Pro is the most actively hostile (to the user) model out of the top 3 (Google, OpenAI, Anthropic). It perfectly reflects Google as a company.
66.
▲
by
tempaccount420
11mo ago
Anycast is hard, which is why only clouds bother with it in the first place :(
67.
▲
by
tempaccount420
11mo ago
> Give it a JSON schema, inject a '{', and sometimes do a bit of custom parsing on the response I would hope that this is not what OpenAI/Anthropic do under the hood, because otherwise, what if one of the strings needs a l
68.
▲
by
tempaccount420
11mo ago
> in the vein of accessibility which has the silly nickname of a11y ironically that's not very accessible...
69.
▲
by
tempaccount420
11mo ago
So you rather stay comfortable doing what you were always doing, which is fine but programming is everything but that...
70.
▲
by
tempaccount420
1y ago
> It feels like an almost deliberate crippling of progress by diverting top talent into useless avenues, much like string theory in physics, or SLS/Artemis. You don't have to be a "top talent" to rewrite old unix util
71.
▲
by
tempaccount420
1y ago
> That being said, I hope they figure out the plugin system, or someone forks the project to add the missing table stakes features. They decided on an obscure Lisp flavor as the language (instead of WASM), so I don't hold my breath
72.
▲
by
tempaccount420
1y ago
The Helix devs have a very specific and unusual taste, so I would never consider PR-ing anything more than a simple bug fix. Not all open-source projects want contributors, many are open-source for different reasons.
73.
▲
by
tempaccount420
1y ago
> Because once I have an intelligence that can actively learn and improve Great news - you already do.
74.
▲
by
tempaccount420
1y ago
Sooo, less reasons (more excuses) for people to move from C++ to Rust?
75.
▲
by
tempaccount420
1y ago
That's usually what happens to these type of forks.
76.
▲
by
tempaccount420
1y ago
Do you find it hard to buy new pants? I wish there was a search engine for clothes...
77.
▲
by
tempaccount420
1y ago
You can get Mac Studio 3rd party "SSDs" for less than half the price Apple charges for the same storage, with the same performance, they even use the same flash chips!
78.
▲
by
tempaccount420
1y ago
Storage-gating is really disgusting considering how much Apple charges for storage.
79.
▲
SSH3: Faster and rich secure shell using HTTP/3
(github.com)
532 points
by
tempaccount420
1y ago
|
263 comments
80.
▲
by
tempaccount420
1y ago
I think that's wukko (either that or both authors are Russian)
81.
▲
by
tempaccount420
1y ago
VS Code is also open source but a lot of Microsoft's extensions for it are closed source. I hope Zed won't go the same path.
82.
▲
by
tempaccount420
1y ago
That's genius.
83.
▲
by
tempaccount420
1y ago
RIP to Gitpod, once a great Remote Development Environment, now a generic AI agent. Couldn't this have been a separate product?
84.
▲
by
tempaccount420
1y ago
> our random spawn point It's not random.
85.
▲
by
tempaccount420
1y ago
WASM should be easier to analyze since you can't look at what functions the WASM imports to do side-effects.
86.
▲
by
tempaccount420
1y ago
kTLS just sounds like a bad idea all around.
87.
▲
by
tempaccount420
1y ago
You really tried to inflict the most amount of damage to the reader by choosing GPT-4o mini
88.
▲
by
tempaccount420
1y ago
Wish they gave us access to learn from those grandmother models instead of distilled slop.
89.
▲
by
tempaccount420
1y ago
It's funny because I was thinking the opposite, the pricing seems way too high for a 5B parameter activation model.
90.
▲
by
tempaccount420
1y ago
You can capture the telemetry data with a HTTPS MITM and read it yourself. Or (if you're working lower level) you can see an obfuscated function is emitting telemetry, saying "User did X", then you can understand that the fun
More ›