Bnjoroge

5 ms·

61.
by Bnjoroge 4mo ago
Nicely done! I've been enjoying using it. Thanks for sharing.
62.
by Bnjoroge 4mo ago
Personally, I disable pre-commit hooks because they’re annoying and slow me down. Pre-push hook would make more sense. Secondly, i’d ideally want to use my codex/ claude subs for this, not an api key
63.
by Bnjoroge 4mo ago
Opencode and other harnesses like Pi do support codex and claude(for now) subscriptions, but the point about security makes sense.
64.
by Bnjoroge 4mo ago
I use OMP. Was on Pi but switched over because I like OMP's subagents implementation. Use 5.5/4.8 for coming up with a very detailed plan, ds4pro/mimo for implementation of easy features, 5.5/4.6 for larger/harder f
65.
by Bnjoroge 4mo ago
How do you a) share context/chat etc with each other? plugin? have them make a file? Why did you opt for either of them instead of something like opencode that supports those models?
66.
Ask HN: What are some good benchmarks for different agent harnesses?
3 points by Bnjoroge 4mo ago|1 comments
67.
by Bnjoroge 4mo ago
I wonder what if anything has changed ever since this article. Is llm-based compression more mainstream?
68.
by Bnjoroge 4mo ago
This post offers virtually no data to back up their objections and reads as LLM-generated for the most part. I
69.
by Bnjoroge 4mo ago
Incredibly cool
70.
by Bnjoroge 4mo ago
I hope they bring it to their apis, especially v4flash. I find myself using mimo 2.5 more since it supports vision and makes it cheap for doing e2e tests with playwright or similar
71.
by Bnjoroge 4mo ago
Have you tried glm 5.2? It’s better than ds4pro and very close to opus4.6. Alot more expensive than ds4pro/mimo but significantly chesper than opus 4.6 for 90% of the quality
72.
by Bnjoroge 4mo ago
You can BYOK with openrouter and pay nothing up until 1mil requests/month iirc which is pretty generous for one person
73.
by Bnjoroge 4mo ago
Deepseek direct is atleast 2x cheaper than other providers serving it. Maybe their caching strategy is just significantly better or they’re subsidizing api pricing. I think it’s the former. $2 is closer to 50 million v4pro tokens in my expe
74.
by Bnjoroge 4mo ago
What are interesting problems? Deepseek and some other open source models do decently well in the gpu kernel benches. https://kernelbench.com/hard Most people think they’re working on “very hard” problems and they really ar
75.
by Bnjoroge 4mo ago
You can specify which providers you want to serve your model in OpenRouter. Then you can chose US-based ones.
76.
by Bnjoroge 4mo ago
They should basically replace any Why Iroh vs Tailscale with this. Thank you
77.
by Bnjoroge 4mo ago
Very cool. I have something similar set up for my homelab where I launch nixos incus containers. Been great for offloading stuff from the my agents
78.
by Bnjoroge 4mo ago
Openrouter has them as significantly more expensive.
79.
by Bnjoroge 4mo ago
I mean you can if you have the compute. They’re all open source
80.
by Bnjoroge 4mo ago
Indeed. I did the math and arrived at the same conclusion. They don’t really subsidize their token plans. Maybe because their api pricing is already dirt cheap
81.
by Bnjoroge 4mo ago
Output tokens are almost 5x more expensive than mimov2.5 pro/dsv4pro. I’m curious to see if Kimik2.7 is that much better. Feels like kimi are positioning themselves as the premium open source models
82.
by Bnjoroge 4mo ago
huh? They all work great in omp/opencode unless you mean their own native clis like kimi code
83.
by Bnjoroge 4mo ago
It depends on what your task is and how precise your prompts are. Planning with fable or 4.8 and laying out the plan in step by step process and coding with mimo v2.5 pro or dsv4pro or qwen 3.7 max and doing a final review with 5.5 has work
84.
by Bnjoroge 4mo ago
I personally dont put any weight to DeepSWE. Other than 5.5 being directionally the best model, it gets the others pretty wrong in my experience. FrontierCode from cognition looks interesting
85.
by Bnjoroge 4mo ago
Yup. I’m hoping this variant fixes these issues.
86.
by Bnjoroge 4mo ago
This is generally been my experience as well, but i think the main reason for claude code being better at understanding intent is their massive system prompt.
87.
by Bnjoroge 4mo ago
They are far from unusable. They aork great for 80-90% of a typical full stack dev. Alot less useful for more noche stuff
88.
by Bnjoroge 4mo ago
it does, think I misunderstood your value prop. best of luck -definitely a real usecase.
89.
by Bnjoroge 4mo ago
love to see jj support, but if I cant use my subs, it's a deal-breaker for me.
90.
by Bnjoroge 4mo ago
plannotator's made me actually want to read my agents' plans. Great product. I've been liking the review as well.

More ›