tosh

12 ms·

421.
by tosh 2mo ago
i'd not be surprised if the current system prompt negatively affects performance at the least it takes away thousands of tokens in the most important part of the context window (!) also see the comment by comboy on contradictions not h
422.
by tosh 2mo ago
what I found noteworthy: early system prompts are a bit more than 300 words, the latest ones 3000+ the opus 5 system prompt has instructions that explain to opus that it might be handling a request that was intended for fable 5: the use
423.
Claude: System Prompts(platform.claude.com)
762 points by tosh 2mo ago|284 comments
424.
Helwan Ha-300(en.wikipedia.org)
1 points by tosh 2mo ago|0 comments
425.
by tosh 2mo ago
https://archive.is/2KQtb
426.
The rise of air-conditioned clothing(ft.com)
4 points by tosh 2mo ago|1 comments
427.
Outperforming cuBLAS on NVFP4(cudaforfun.substack.com)
1 points by tosh 2mo ago|0 comments
428.
Dario Amodei on Regulation(twitter.com)
4 points by tosh 2mo ago|0 comments
429.
Notes on building products people love when software gets cheap(twitter.com)
2 points by tosh 2mo ago|0 comments
430.
Oman says oil spill reaches coastline(aljazeera.com)
3 points by tosh 2mo ago|0 comments
431.
Relentlessly Resourceful (2009)(paulgraham.com)
1 points by tosh 2mo ago|0 comments
432.
Is AI Coding Effecting Our Mental Health? [video](youtube.com)
1 points by tosh 2mo ago|0 comments
433.
by tosh 2mo ago
I also noticed the ui/ux of cloudflare got way better in the last weeks. That's a good sign. Usually ui/ux in large companies gets worse not better!
434.
by tosh 2mo ago
I think like with most large companies that still ship new stuff: it gets a bit more complex for customers to figure out what the good nuggets are and what to ignore for now same @ google, aws and so on (I usually rely on opinions of people
435.
VictoriaMetrics: AI Policy(docs.victoriametrics.com)
1 points by tosh 2mo ago|0 comments
436.
by tosh 2mo ago
Training material seems to be especially rich re GPU kernels and SIMD. I wonder if there is extra effort put into this because they are useful for the researchers working on the models or just a sub-domain that language models are a great f
437.
by tosh 2mo ago
high leverage for most agents: - review system prompt + cut it down or remove completely - review agents.md file(s), check which ones are loaded, remove or improve them - review context spam from tools, skills etc, de-activate all,
438.
Auto-research with codex: How I achieved a 232x Faster Kernel(sankalp.bearblog.dev)
457 points by tosh 2mo ago|93 comments
439.
A Swarm of Blood Robots(craigmod.com)
1 points by tosh 2mo ago|0 comments
440.
Google Python Style Guide(google.github.io)
1 points by tosh 2mo ago|0 comments
441.
Trump Orders Navy to Restore Older Tech on Aircraft Carriers, Costing Billions(wsj.com)
5 points by tosh 2mo ago|0 comments
442.
No Plan Survives Contact with the Enemy (Reality)(blog.exe.dev)
1 points by tosh 2mo ago|0 comments
443.
The pursuit of excellence does not need justification(twitter.com)
1 points by tosh 2mo ago|0 comments
444.
Omarchy v4.0.0(github.com)
4 points by tosh 2mo ago|0 comments
445.
Qwen 3.8 27B in Cloudflare OS made me a collaborative whiteboard(twitter.com)
1 points by tosh 2mo ago|0 comments
446.
GLM-5.3: How Chinese labs keep stride with the frontier(interconnects.ai)
4 points by tosh 2mo ago|0 comments
447.
Tankers Are Going Dark for Longer to Get Oil Out of Middle East(bloomberg.com)
2 points by tosh 2mo ago|0 comments
448.
"That's not SOC 2 compliant"(ampcode.com)
52 points by tosh 2mo ago|46 comments
449.
by tosh 2mo ago
afaiu this is a kdb-adjacent columnar runtime and query engine think SIMD, morsels and so on ~ Clickhouse, DuckDB … but more like a lisp-like vector language and execution engine at its core + allowing for numeric and analytics adjacent use
450.
Visual Studio Code 1.133(code.visualstudio.com)
3 points by tosh 2mo ago|0 comments

More ›