swingboy

5 ms·

31.
by swingboy 2mo ago
> The inflated price you pay at gas station, groceries, and in interest when you're borrowing money, is a result of those broken promises. Citation needed.
32.
by swingboy 2mo ago
Does Artificial Analysis use OpenRouter for model access to do their benchmarks?
33.
by swingboy 2mo ago
A lot of private repos are probably just taking up cheap storage space, not seeing activity that would cause additional load.
34.
by swingboy 2mo ago
How detailed of a plan? Are you including code snippets or just behavior and letting the lesser model decide how to implement?
35.
by swingboy 2mo ago
Pretty sure Anthropic and other providers prepend these "official" system prompts to your conversation even if you send in a custom system prompt otherwise it would be trivial to produce CSAM, etc.
36.
by swingboy 2mo ago
I think it depends on your subscription? I just have the 20$ and the cache is only 5 minutes. I’ve got the timer in my status line via ccstatusline and unless it is wrong, it says 5 minutes.
37.
by swingboy 2mo ago
How about Anthropic just be more generous with their usage limits instead?
38.
by swingboy 2mo ago
I’ve found Flash 0731 to be pretty great recently. I do feel like a lot of the models are pretty close in terms of ability. I often run code through multiple different models _and_ harnesses for code reviews and they all typically find the
39.
by swingboy 2mo ago
The output of LLMs is already humanized…
40.
by swingboy 2mo ago
What?
41.
by swingboy 2mo ago
Still paywalled.
42.
by swingboy 2mo ago
Agreed. With less than $10 on the DeepSeek API used, I’m somewhere near half a billion tokens over the past week or however long it’s been since it came out. I’ve found it to be very capable. I’m using it with pi as well and some custom ext
43.
by swingboy 2mo ago
Great looking keyboard but why are the numbers on the number keys so large? They stick out like a sore thumb.
44.
by swingboy 2mo ago
Aside from the minimal system prompt, how does it handle context better than other agents? It still has to send the system prompt (which includes AGENTS.md and skill definitions) along with the full conversation every request, no?
45.
by swingboy 2mo ago
Sir, you dropped your fedora.
46.
by swingboy 2mo ago
This is awesome. Luna is a pretty great model on xhigh.
47.
by swingboy 2mo ago
Say a single Kimi K3 is deployed on 16 x B200s: how many concurrent users can that handle? I realize the question assumes a major simplification that everyone's prompts/sessions are the same.
48.
by swingboy 2mo ago
What is this writing style…OP is a 12 year old girl? Penguin of doom…
49.
by swingboy 2mo ago
Yet they do.
50.
by swingboy 3mo ago
Refilling the context of the cheap executor == just switching the model mid-conversation instead of /clear’ing and passing some plan doc?
51.
by swingboy 3mo ago
On the first syntax example: there’s something funny to me about using three pipe operator and four different functions to turn “hello world” into “Hello World”.
52.
by swingboy 3mo ago
This doesn’t look as fun as a Tamagotchi.
53.
by swingboy 3mo ago
What are some of the things you’re doing with the Codex app-server?
54.
by swingboy 3mo ago
The `repo_path` field.
55.
by swingboy 3mo ago
Well, it looks like he was running the agent in his home directory to begin with considering the `repo_path` field is exactly that.
56.
by swingboy 3mo ago
Is it because Pi’s default system prompt is so simple?
57.
by swingboy 3mo ago
Haven’t heard this one in a while XD
58.
by swingboy 3mo ago
Which 2 or 3 skills?
59.
by swingboy 3mo ago
There was a big thread about it here the other day. https://news.ycombinator.com/item?id=48734373
60.
by swingboy 3mo ago
Serious question in good faith: what was the deal with the “calamari” (clots?) the anti-vax crowd kept talking about being found in the veins/arteries of folks who took the Covid vaccine?

More ›