Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
paxys
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
12 ms
·
421.
▲
by
paxys
4mo ago
> AI is making this all so much worse. When you are prompting you feel like you are steering, but are you really? Would you know if you weren’t? This one hits hard. I feel more and more that AI-assisted creation is really just consumptio
422.
▲
by
paxys
4mo ago
Hires like these are made for their brand rather than expertise. His job will be to attend galas, schmooze with investors and give legitimacy to the operation. The company will hire others to do the real work.
423.
▲
by
paxys
4mo ago
I've been getting "this model is at capacity" a bunch with Sol, so definitely launch related.
424.
▲
by
paxys
4mo ago
Macs have a single digit market share globally, so even if people started installing games on them the needle wouldn't move much.
425.
▲
by
paxys
4mo ago
Facebook was a networking site for a handful of college campuses at the time. It had some great traction, sure, but I doubt "filthy rich" was a serious expectation among employees.
426.
▲
by
paxys
4mo ago
Whatever Zuck is feeling on any given day. At the moment it seems to be prediction markets .
427.
▲
by
paxys
4mo ago
If that was the case then he wouldn't have continued doubling and tripling down on these same product decisions for many more years. Ask anyone at Meta and they will readily agree that Boz is among the most incompetent and unserious le
428.
▲
by
paxys
4mo ago
Takes some extraordinary lack of self awareness to write something like this while burning tens of billions of dollars a year spearheading the "Metaverse", which is as big a digression from the company's core competency as yo
429.
▲
by
paxys
4mo ago
Where's your website?
430.
▲
by
paxys
4mo ago
I'm just happy there's competition.
431.
▲
by
paxys
4mo ago
I have found that higher reasoning doesn't always produce better results. Models often tend to overthink and get themselves in a bind. And tasks just take longer for no reason.
432.
▲
by
paxys
4mo ago
Uber is already “Uberlance”. Most people take Ubers to the hospital unless they are literally dying (and some even then).
433.
▲
by
paxys
4mo ago
But that’s socialism!
434.
▲
by
paxys
4mo ago
Big difference between the two - America has elections.
435.
▲
by
paxys
4mo ago
Which is why the combined app is called ChatGPT, not Codex.
436.
▲
by
paxys
4mo ago
Doesn’t really matter what it stands for. ChatGPT is one of the most recognizable brands in the world right now.
437.
▲
by
paxys
4mo ago
It’s all ChatGPT under the hood. Even though it says Work or Codex or whatever you can just ask it whatever you want, and it will work.
438.
▲
by
paxys
4mo ago
The site they showed had the pelican riding a number of things, including a bicycle and another pelican (no, not in that way).
439.
▲
by
paxys
4mo ago
Okay Richie Rich
440.
▲
by
paxys
4mo ago
Sam Altman confirmed during the initial limited release that Sol will run on Cerebras at 750 tok/sec.
441.
▲
by
paxys
4mo ago
Rebranding + expanding it beyond coding work.
442.
▲
by
paxys
4mo ago
Feels like forever in the AI world but Claude Cowork only launched ~5 months ago. So not that overdue.
443.
▲
by
paxys
4mo ago
Makes sense why they released an entire study yesterday discrediting SWE-bench Pro.
444.
▲
by
paxys
4mo ago
Where’s the lie?
445.
▲
by
paxys
4mo ago
How is it their first try? They were leading the race with Llama 3.x a few years ago.
446.
▲
by
paxys
4mo ago
How is every company able to show itself at the top of every benchmark?
447.
▲
by
paxys
4mo ago
Why not wait one more day for GPT-5.6?
448.
▲
by
paxys
4mo ago
SWE-Bench Pro was created to replace SWE-Bench and fix these problems.
449.
▲
by
paxys
4mo ago
My vesting cliff somehow hit after getting fired lol
450.
▲
by
paxys
4mo ago
They picked Go after meaningfully considering Rust (and others). I don't remember all the reasons for it but it was detailed in the original blog post.
More ›