Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
golly_ned
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
8 ms
·
91.
▲
by
golly_ned
1y ago
Twice so far, though not in as stark terms, and not recently. Once, I told him to knock it off trying to make a peer engineer who was underperforming look bad in front of the team, and instead to inform our manager of cases where he believe
92.
▲
by
golly_ned
1y ago
I think from what I've written, it could be possible to interpret this as the case. At least part of it's true -- I'm not nearly as experienced in some of the technologies this team uses as the other engineer who's been
93.
▲
by
golly_ned
1y ago
Thank you for your advice. > your manager isn’t doing his job Yeah, I do believe this and agree. My manager's new, and hasn't gotten much guidance or mentorship. I feel like he depends on me for a lot, and since I don't ha
94.
▲
by
golly_ned
1y ago
This sounds similar to the guy I'm working with. He's great in certain ways and takes his job seriously but can be really grating to work with. (Frankly, I think it probably has something to do with taking stimulants, just based o
95.
▲
Ask HN: Not treated respectfully by colleague – advice?
116 points
by
golly_ned
1y ago
|
128 comments
96.
▲
by
golly_ned
1y ago
The syntax is based on python, but its runtime is not. So nothing about the contrast between the python language and mojo's use as a super-parallelized parallel processing system is inconsistent.
97.
▲
by
golly_ned
1y ago
Tritonlang itself is a deep learning DSL.
98.
▲
by
golly_ned
1y ago
Amazon did this, at least for a time, as part of a "No See, No Hear" hiring pilot program. The purpose was to see if they could hire university graduates with a minimum of human interviewing effort. They selected from a handful of
99.
▲
by
golly_ned
1y ago
> "The technology at issue was among the most transformative many of us will see in our lifetimes" A judge making on a ruling based on his opinion of how transformative a technology will be doesn't inspire confidence. Ther
100.
▲
by
golly_ned
1y ago
Of all the languages one could accuse of being hermetically designed in an ivory tower, Go would be the second-least likely.
101.
▲
by
golly_ned
1y ago
You'd have to commit yourself to believing a massive amount of implausible things in order to address the remote possibility that AI consciousness is plausible. If there weren't a long history of science-fiction going back to the
102.
▲
by
golly_ned
1y ago
I take good care of my pet rock for the same reason. In case it comes alive I don't want it to bash my skull in.
103.
▲
by
golly_ned
1y ago
Speaking from my experiences in dozens of debriefs: I don't think I can recall a case where an otherwise good candidate was rejected for something like not liking someone's personality. There might be a middle ground between "
104.
▲
by
golly_ned
1y ago
It strikes me as a massive anti-pattern to have one developer be the sole contributor of an open-source project sponsored by a $3T company. It doesn't speak well to its longevity or the strength of the sponsorship Microsoft's putt
105.
▲
by
golly_ned
1y ago
Derek Thompson is not well suited to this kind of work. He is much better suited to his usual lame, predictable, centrist, lukewarm commentary on tired political and social topics.
106.
▲
by
golly_ned
1y ago
Are there any such build-a-lisp books/guides using modern c++?
107.
▲
by
golly_ned
1y ago
I have a similar use case and this idea also occurred to me. However: the dict in this case would also include dataclasses, and I’d be interested in finding what exact attributes within those dataclasses were accessed, and also be able to m
108.
▲
by
golly_ned
1y ago
It often invents sources. At least for me.
109.
▲
by
golly_ned
1y ago
I still don’t quite get the motivation for “don’t use pydantic except at border” — it sounds like it’s “you don’t need it”, which might be true. But then adds dacite to translate between pydantic at the border and python objects internally.
110.
▲
by
golly_ned
1y ago
Considering the human brain evolved over hundreds of millions of years of evolution (what a lot of data!) that almost seems cheap to do in the span of a number of months to train a model.
111.
▲
by
golly_ned
1y ago
Those are compute credits that are directly spent on the experiment itself. It's no more "compensation" than a chemistry researcher being "compensated" with test tubes.
112.
▲
by
golly_ned
1y ago
[flagged]
113.
▲
by
golly_ned
1y ago
The guy who was disciplined for reading poetry on duty during WWII, who threw away his commissioned abstract sculpture would feel perfectly comfortable on a forum clearly hostile to anything that isn't a hard fact or a short thinkpiece
114.
▲
by
golly_ned
1y ago
How was the comment ableist?
115.
▲
by
golly_ned
1y ago
What a beautiful article about a fascinating man -- I'd be glad to read a full biography.
116.
▲
by
golly_ned
1y ago
"Cancer is cured, the economy grows at 10% a year, the budget is balanced — and 20% of people don't have jobs" That's a "prediction" by Amodei. It's just words. Pure bullshit. Diminishing my otherwise high
117.
▲
by
golly_ned
1y ago
It's practically unlimited usage with $100 or $200 subscription. Given prices elsewhere now I don't see how this is viable longer-run and not heavily subsidized by losses in anthropic
118.
▲
by
golly_ned
1y ago
The author is reknowned for eMacs lisp.
119.
▲
by
golly_ned
1y ago
There’s a section that says what the parts of a pc are, and what that part is.
120.
▲
by
golly_ned
1y ago
Whenever I get to a section that was clearly autogenerated by an LLM I lose interest in the entire article. Suddenly the entire thing is suspect and I feel like I’m wasting my time, since I’m lo lingering encountering the mind of another pe
More ›