Xing Xiang
@_xingxiang
Followers
57
Following
1K
Media
7
Statuses
305
Software engineer | Building https://t.co/hdKBndxNDA for medical students (LLM in medical education)
Singapore
Joined January 2021
RT @leerob: What the heck is "Qwen3-Coder-480B-A35B-Instruct"?. It sounds like a series of random numbers and letters. But it’s not too dif….
0
44
0
Smart! Forcing the model to choose a subset of tools using naming and message prefilling , without repeating the tool definitions in the latest message. Also great point on optimising prompts around KV cache for speed + cost
After four overhauls and millions of real-world sessions, here are the lessons we learned about context engineering for AI agents:
0
0
0
RT @softwaredoug: The older and dumber I get, the more I just want some grug-brained evals to test search changes. Instead of a perfect….
softwaredoug.com
Big brain spend months building perfect quality metrics. Grug brain no trust, and just want dumb labels from coworkers 👍/👎.
0
2
0
RT @jeffwsurf: To put it mildly, the past week at Windsurf has been crazy. There have been a lot of different rumors and reports, so I want….
0
249
0
RT @thenanyu: Compensation curves for IC engineers at *big companies* has always been incorrect in my view. What Zuck is doing right now is….
0
106
0
RT @HongLiu9903: 5 / 6. The most exciting implication: We can now continuously feed massive, unlabeled web-scale multimodal data into MoCa….
0
1
0
RT @Suhail: PSA: there’s a guy named Soham Parekh (in India) who works at 3-4 startups at the same time. He’s been preying on YC companies….
0
3K
0
RT @HuggingPapers: Forget everything you thought about LLM prompting!. New research introduces PromptQuine, a self-replicating framework th….
0
3
0
RT @PlanetScale: PlanetScale is the fastest Postgres provider. See how we stack up in benchmarks:.
planetscale.com
Benchmarking Postgres in a transparent, standardized and fair way is challenging. Here, we look at the process of how we did it in-depth
0
7
0
RT @Xianbao_QIAN: How fast is LLM evolving? GPT-4o last year is caught up by a 9B visual thinking model now. The new model from @thukeg se….
0
50
0
RT @dillon_mulroy: how i write effect (Effect.gen vs .pipe). i use Effect.gen and yield* for business logic and .pipe for composition and s….
0
23
0
RT @_hiteshbandhu: here are my takeaways from a recent @aiDotEngineer video on telemedicine agents. @swyx . [the video link is available in….
0
5
0
RT @harrysolovay: Use Effect AI to embed chunks of the novel Dracula. Store those chunks in a Postgres database using pgvector and the Eff….
0
10
0
RT @mattpocockuk: Here's how to guarantee you've handled every known error in your system. from the very top of a potentially enormous call….
0
24
0
RT @bo_wangbo: jina-embeddings-v3 + jina-clip-v2 + jina-colbert-v2 + colpali + dse = jina-embeddings-v4 😇. https://….
0
32
0
RT @jxnlco: why “bother” fine-tuning an embedder?.it can be powerful - but it’s not for everyone. fine-tuning an embedding model can raise….
0
10
0
RT @emollick: New paper shows a familiar result on LLMs & medicine:. Doctors given clinical vignettes produce significantly more accurate d….
0
197
0
RT @emollick: 🚨We have a new prompting report:. Prompting a model with Chain of Thought is a common prompt engineering technique, but we fi….
0
58
0
RT @leerob: I'm using agents to write a lot of code now. Just 6 months ago it didn't really work well, so if you haven't tried in a while,….
0
49
0