Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Havoc
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
121.
▲
by
Havoc
2mo ago
The notion of AI having an equalizing effect seems absurd to me. Buying tokens is about as pay-to-win as it gets. More so than anything we’ve seen in a long time. Before everyone had the same 24hr in a day which despite inequalities grounde
122.
▲
by
Havoc
2mo ago
That is true but “is never seen in the workplace” to doesn’t add value is still one hell of a leap. At scale CEOs don’t add value by clocking in and pumping out LoC themselves and they sure aren’t having all their employees tag along on the
123.
▲
by
Havoc
2mo ago
Not really. I’m not contrasting two peoples opinions on the same thing, but rather aggregate vibe on two different but functionally similar things. Bit like if hn was convinced white cats catch mice better than black ones. Point out that t
124.
▲
by
Havoc
2mo ago
Neither CEO nor dev here but must say hn comments on CEOs are always wild. It’s like some caricature - like an attempt to draw an animal that has never been seen from a description Somehow founders constantly putting out fires, leading and
125.
▲
by
Havoc
2mo ago
Their list of allowed tools is extensive so just use whatever you want within that list Think z code gives a token bonus though
126.
▲
by
Havoc
2mo ago
> Big ole pool of very fast ram It’s more that it’s a very parallel architecture than fast. The LPDDR5X under the hood is slower than what you’d find in a graphics card vram of years ago That’s why you get consumer macs with 512gb while
127.
▲
by
Havoc
2mo ago
You’ll be waiting a while. Ram capacity is bought out for 2027
128.
▲
by
Havoc
2mo ago
And twitters gotten increasing aggressive in forcing you to download the app and log in Irritates me every time there is a hn link to tiwtter
129.
▲
by
Havoc
2mo ago
Has yet to materialise
130.
▲
by
Havoc
2mo ago
40m sounds quite heavy for a qwen fine tune?
131.
▲
by
Havoc
2mo ago
> and all previous models They have had vision models before just not their flagships
132.
▲
by
Havoc
2mo ago
Can’t see myself buying my this but glad riscv is making progress.
133.
▲
by
Havoc
2mo ago
I need my porch painted and tradesmen are expensive
134.
▲
by
Havoc
2mo ago
The Irish example seems good. Keeping the old tech in parallel rather than replacing
135.
▲
by
Havoc
2mo ago
They’re just lining up the Xth once in a lifetime crisis for millennials…
136.
▲
by
Havoc
2mo ago
The story is about the British system not the EU one
137.
▲
by
Havoc
2mo ago
One of the reasons I’m applying for a British passport. Existing purely in some mystery database like that seems risky long term
138.
▲
by
Havoc
2mo ago
Intentionally going after old people because they’re more susceptible to AI and declaring as much on record is fuckin wild. Does SV now optimize for evilness?
139.
▲
by
Havoc
2mo ago
Wow that demo video is way better than expected. Well done
140.
▲
by
Havoc
2mo ago
Door dash
141.
▲
by
Havoc
2mo ago
That would be an essay, but in short: Grafana is fine to get to a selfhosted basic install. But once you try to actually connect logs, metrics, traces in selfhosted context and perhaps sprinkle some otel in...that sht gets out of hand very
142.
▲
by
Havoc
2mo ago
I find the entire observability space to quite a poor experience, at least in the self-hosted space. Tried both grafana route and signoz and neither seems particularly pleasant
143.
▲
by
Havoc
2mo ago
>you can choose how many bugs you want now, You are a rust expert. Make only 5 mistakes. :) Interesting quote though - I guess one does get to decide how much testing/debugging effort one wants now
144.
▲
by
Havoc
2mo ago
From afar the US is really sounding quite scary lately
145.
▲
by
Havoc
2mo ago
That’s quite a feat
146.
▲
by
Havoc
2mo ago
Complete nonsense. > they provide a better or at least as good an answer as LLMs in 62.5% of the cases. Are we going to scrap hospitals because a vet could do the job 62.5% of the time? The economics also point away from everyone buying
147.
▲
by
Havoc
2mo ago
Google doing anti competitive bullshit? Never
148.
▲
by
Havoc
2mo ago
Could have sworn I read this exact article a month ago
149.
▲
by
Havoc
2mo ago
Meanwhile I can’t get a western LLM to look at a repo and tell me whether it contains anything malicious (it was a skill repo - literally just text files). Alignment my ass
150.
▲
by
Havoc
2mo ago
Remember when gpt2 was too dangerous to release? Something being dangerous and sama saying something is dangerous are not necessarily the same thing. Especially when he’s got everything riding on this bet
More ›