Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
spenczar5
searching Neon…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
13 ms
·
1.
▲
by
spenczar5
9d ago
> What happens when some stick-up-the-ass company—whose stuff is probably in hundreds of different collections—decides it really wants a payday? It's happened: - https://en.wikipedia.org/wiki/Hachette_v._Interne
2.
▲
by
spenczar5
11d ago
This is still the case today, if you let it be. I am friends with many neighbors with differing views from my own and literally do have them over for dinner, invite them to my toddlers' birthday parties, etc. I don't recognize t
3.
▲
by
spenczar5
2mo ago
> This is still $150M of pure gravy every single year. The issue is rate of return. They are evidently spending $4.85bn on Xbox per year. The US federal interest rate is 3.5% so if you just put that money in US bonds you would get about
4.
▲
by
spenczar5
4mo ago
Sometimes you get there by accident. You make a thing, it grows or is used in unexpected ways, now suddenly performance matters. Sometimes Python is just the language used in the domain. Lots of sciences live on Python because it is easy to
5.
▲
by
spenczar5
4mo ago
It's in the first sentence of your quote: "our continued collaboration with Anthropic" Read this as: "we get discounts, rate limit increases, a direct line to responsible product managers; in exchange we participate in f
6.
▲
by
spenczar5
4mo ago
Yep! The industry term is "co-marketing" and its hard to avoid seeing once you spot it.
7.
▲
by
spenczar5
4mo ago
No, but they go on strike when negotiating their collective contracts, and put terms in the contract that govern how failures like this are investigated and punished.
8.
▲
by
spenczar5
4mo ago
> A correction was made on May 6, 2026: An earlier version of this article misstated the country where Guy Goma grew up. He is from the Republic of Congo, not the Democratic Republic of Congo. Right guy, wrong Congo! You can't even
9.
▲
Agent Engineering Patterns: Dealing with large tool results
(blog.firetiger.com)
3 points
by
spenczar5
6mo ago
|
0 comments
10.
▲
by
spenczar5
6mo ago
Does it matter that pen and paper dominate? How much of the business's expenses are overhead?
11.
▲
by
spenczar5
6mo ago
Sure, llms.txt is a convention for this. Compare https://docs.firetiger.com with https://docs.firetiger.com/llms.txt and https://docs.firetiger.com/llms-full.txt for a realy example.
12.
▲
by
spenczar5
6mo ago
That's a pretty interesting idea! I guess 160+ is sort of doing some of that for us - it compiles to SQL WHERE clauses, right - but generally, we found good results giving it a SQL dialect directly. I think some of the reason is that
13.
▲
by
spenczar5
6mo ago
Yes! This works really well from Sonnet 4.5 onwards, in our experience. Sonnet 4.0 was a little rocky - we had to give it tons of documentation - but by now it works without much effort. One thing that works very well is just giving it one
14.
▲
by
spenczar5
6mo ago
Author here! I am pretty jazzed about these ideas and happy to dig into more detail than a blog post allows.
15.
▲
by
spenczar5
6mo ago
Is this a clone of the Google AIPs? Like https://aep.dev/160/ seems to just copy https://google.aip.dev/160 .
16.
▲
by
spenczar5
7mo ago
"Cheap" how? I have a friend who works on Seattle's bus planning. Removing a stop is a _lot_ of political work. When an elderly person depends on that bus stop being within a block so they can get to their doctor, and you
17.
▲
Dynamically making agents to monitor PRs
(blog.firetiger.com)
2 points
by
spenczar5
7mo ago
|
0 comments
18.
▲
by
spenczar5
7mo ago
Its unexported for that reason. You only change it in tests.
19.
▲
by
spenczar5
8mo ago
It's in the article that you're commenting on, https://www.spacex.com/updates#xai-joins-spacex .
20.
▲
by
spenczar5
9mo ago
I feel like I see an independent low-noise phone project like, every 3 months. Clearly there is some latent demand here. I wonder why the big players (Google, Apple, Samsung, HTC) haven't made a big-corp product for this market. I am a
21.
▲
by
spenczar5
10mo ago
"But accepting the full S3Client here ties UploadReport to an interface that’s too broad. A fake must implement all the methods just to satisfy it." This isn't really true. Your mock inplementation can embed the interface, bu
22.
▲
by
spenczar5
10mo ago
no, "pdf" is a very typical shortening for "probability density function," its correct.
23.
▲
by
spenczar5
10mo ago
Its JSON schema, well standardized, and predates LLMs: https://json-schema.org/
24.
▲
by
spenczar5
11mo ago
Thanks for the detailed answer! I use DuckDB today to query Iceberg tables. In some particularly gnarly queries (huge DISTINCTs, big sorts, even just selects that touch extremely heavy columns) I have sometimes run out of memory in that Duc
25.
▲
by
spenczar5
11mo ago
Very cool. One question that comes up for me is whether pg_lake expects to control the Iceberg metadata, or whether it can be used purely as a read layer. If I make schema updates and partition changes to iceberg directly, without going thr
26.
▲
by
spenczar5
11mo ago
I think I don't understand postgres enough, so forgive this naive question, but what does pushing down to the remote tables mean? Does it allow parallelism? If I query a very large iceberg table, will this system fan the work out to mu
27.
▲
by
spenczar5
11mo ago
sure, but that is a fairly trivial tool call too. Ask it to name the distribution family and its parameter values.
28.
▲
by
spenczar5
11mo ago
Frankly, yes. The models are one part of the story. But the software around it matters at least as much: what tools does the model have access to, like bash or just file reading or (as in your example!) just a cache of files visited by the
29.
▲
by
spenczar5
1y ago
Where (vaguely, of course) do you live? That sounds so different from my experience in a city.
30.
▲
by
spenczar5
1y ago
I agree, but there are other possibilities in between those two extremes, like Quivr [1]. Schemas are good, but they can be defined in Python and you get a lot more composability and modularity than you would find in SQL (or pandas, realist
More ›