Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
phonon
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
4 ms
·
1.
▲
by
phonon
7d ago
I don't think anything will beat https://www.asahi.com/sp/ajw/articles/15148829 School principal fired and loses $130k pension for overfilling coffee (about 50 cents).
2.
▲
by
phonon
12d ago
"Officer" is typically anyone who can bind the company contractually.
3.
▲
by
phonon
15d ago
Previously https://news.ycombinator.com/item?id=48436863
4.
▲
by
phonon
19d ago
The entire argument rests on a 36-year-old survey form: "DOE Form 860 (1990) defines electric utility to mean 'a corporation, person, agency, authority, or other legal entity or instrumentality that owns and/or operates facil
5.
▲
by
phonon
21d ago
How is this different than https://foundershield.com/ ?
6.
▲
by
phonon
25d ago
Uhhh, we're not doing so hot there either. https://www.propublica.org/article/general-dynamics-artiller...
7.
▲
by
phonon
27d ago
https://news.ycombinator.com/item?id=47556729
8.
▲
by
phonon
1mo ago
OpenAI (Stargate) kicked off the price increases a year ago by locking up 40% of the global supply of RAM output. https://www.tomshardware.com/pc-components/dram/openais-star...
9.
▲
by
phonon
1mo ago
A few hundred bad bits/transistors in a trillion+ parameter model would compromise its abilities not one iota...the models are inherently lossy and resistant to "brain damage"...
10.
▲
by
phonon
1mo ago
A full wafer like Cerebras is about 60x that, and N2P has about 3x the transistor density. So right now it's technically feasible to etch a 1.4 trillion parameter model. So roughly DeepSeek-V4-Pro class. Imagine that running a factory
11.
▲
by
phonon
1mo ago
Because Trump hates wind power for "ruining the view". That's it. https://www.youtube.com/shorts/uyLiON_f5c0 https://www.bbc.com/news/uk-scotland-north-east-orkney-shetl...
12.
▲
by
phonon
1mo ago
Are they conjoined triangles?
13.
▲
by
phonon
2mo ago
Pratt & Whitney PW2040 turbofan engine, same as the Boeing 757 and the C-17. They say that the plane itself is so efficient they don't need a cutting edge efficient engine.
14.
▲
by
phonon
2mo ago
Codex context limit is 400k total, including 128,000 reserved output tokens.
15.
▲
by
phonon
2mo ago
Global Foundries has lots of fabs with 10 year old+ processes....
16.
▲
by
phonon
3mo ago
M3 Ultra has a 1024 bit memory bus (819 GB/s) and starts at $3,999 (96GB of RAM). It can be done....
17.
▲
by
phonon
3mo ago
How does it compare to https://news.ycombinator.com/item?id=48436863 ?
18.
▲
by
phonon
4mo ago
Ticket quotas.
19.
▲
by
phonon
4mo ago
Use EMV 3DS 2.x authentication with liability shift protection?
20.
▲
by
phonon
4mo ago
"... this flying wing will burn 50% less fuel than today's jets..."[1] [1] https://time.com/7292452/jetzero-low-carbon-air-travel/
21.
▲
by
phonon
4mo ago
It would depend on exactly how the records were obtained. If hypothetically she paid a nurse $100 for it (and had done so many times in the past), that could be criminally liable. In practice, no-one would prosecute over a single x-ray. But
22.
▲
by
phonon
4mo ago
If you induce someone to violate HIPAA who is covered by it (like say a nurse at a hospital), you can be criminally liable. There is no carve-out for journalists. BOTH the person who gave the record and the person who induced them to give i
23.
▲
by
phonon
4mo ago
A. That's how I read it too. B. You can be criminally liable for HIPAA violations, if you induce someone covered by them to violate them. See for example https://www.justice.gov/usao-nj/media/1254226/dl
24.
▲
by
phonon
4mo ago
https://www.codereviewbench.com/ https://codereview.withmartian.com/
25.
▲
by
phonon
4mo ago
When it's ICE it's both :-(
26.
▲
by
phonon
5mo ago
That's coming out in https://www.phoronix.com/review/intel-crescent-island by around the end of the year.
27.
▲
by
phonon
5mo ago
Microsoft has https://blogs.microsoft.com/blog/2026/01/26/maia-200-the-ai-... already.
28.
▲
by
phonon
5mo ago
> Acetylcysteine That's NAC (N-acetylcysteine, C5H9NO3S), mentioned in the article many times.
29.
▲
by
phonon
5mo ago
4 DIMMS =/= 4 channels
30.
▲
by
phonon
5mo ago
Because Cerebras handles large models poorly due to latency/bandwidth issues to main memory. See https://openai.com/index/introducing-gpt-5-3-codex-spark/ where its performance is significantly below that of
More ›