HackerTrans
TopNewTrendsCommentsPastAskShowJobs

ammar_x

no profile record

Submissions

DeepSWE: A contamination-free benchmark for long-horizon coding agents

deepswe.datacurve.ai
67 points·by ammar_x·2 mesi fa·20 comments

Xiaomi Mimo-v2.5 pricing is now permanently reduced

twitter.com
5 points·by ammar_x·2 mesi fa·0 comments

comments

ammar_x
·2 mesi fa·discuss
Is there some sort of a leaderboard for this test? Like if you'd give each of Opus 4.8 and GPT 5.5 a score out of 100, what would the scores be?
ammar_x
·2 mesi fa·discuss
Absolutely! We need new and better benchmarks like this.

I have a question: why not use the maximum available reasoning on each LLM? For example, I see that Opus 4.7 at `max` reasoning but Sonnet 4.6 at `high`. Wouldn't it be a fairer comparison if all were at max?
ammar_x
·2 mesi fa·discuss
https://x.com/serenaa_ge/status/2059308400866111692
ammar_x
·2 mesi fa·discuss
I usually do this for complex features:

- Opus 4.7 writes the code - I make GPT-5.5 in Codex to review it (given context) - I provide the review back to Opus and ask it to verify the review findings - Make Opus plan the fixes then execute them - Ask GPT-5.5 to review the fixes and check if they solve the problems
ammar_x
·2 mesi fa·discuss
Cool, but body font size is too small for comfortable reading!
ammar_x
·2 mesi fa·discuss
I've used Brave Search and found it better than Google's in some cases
ammar_x
·2 mesi fa·discuss
This looks great for quick audio operations without the need to use heavy apps.

One question: I tried the "Fade In" effect; is there a way to control its timing (i.e. the part of the clip where the effect is applied) ?
ammar_x
·2 mesi fa·discuss
You can use V4 Pro with Claude Code [1].

I tried it and it's impressive.

[1]: https://api-docs.deepseek.com/quick_start/agent_integrations...
ammar_x
·9 mesi fa·discuss
My "trick" was to divide things into batches (which can be big with LLMs with larger context sizes) and classify the items in each batch, then take the resulting categories from each batch and feed them into an LLM to group semantically similar categories into groups with a representative category for each group. The representative category can be chosen from the group or created by the LLM. This is an over-simplification of the process but that's the gist of it.
ammar_x
·9 mesi fa·discuss
Language support is not mentioned in the repo. But from the paper, it offers extensive multilingual support (nearly 100 languages) which is good, but I need to test it to see how it compares to Gemini and Mistral OCR.
ammar_x
·9 mesi fa·discuss
Claude Skills seem to be the option that offers highest flexibility to add more capabilities at most simplicity. Better than MCP in my opinion. Hope it becomes a standard and get adopted by OpenAI and the rest of labs.