Sabitlenmiş Tweet
Ashish
958 posts
@_ashleypeacock It’s next evolution is “Let me investigate and circle back”
English
Kimi K2.6 feels like a true software engineer with the constant "Hmm..." and "Wait..." while thinking
English
Ashish retweetledi
Yesterday Workers AI rolled out two major changes for Kimi k2.6: speculative decoding and shared KV cache. The first should improve TPS, especially for predictable output like tool calls. The second should improve cache hit rates, lowering TTFT and blended cost.
English
If everyone had a product owner who cared about their product even just 10% of what @michellechen does, the world would be a better place.
michelle@michellechen
couldn't be possible without these big brains: ✨ Workers AI team - @kevin_flansburg @MichaelDwan @mark_a_phelps @kevinjain_ @xjewer, lane, dan 🤖 MLE team - @uhsheeeesh @kevin_flansburg @jtkipp, piotr, dennis. and @lucatac0 @shridharathi @allnoteson who helps us test BYO model work 🔮AI Gateway team - @minglu @superhighfives @mattrothenberg @_irtefa, sam, joaquin, gandalf, nikhil, nikita, matt
English
@michellechen @davzaman I just played a game of league. I can confirm I am a bot.
English
Ashish retweetledi
kimi is here. workers ai is officially in the big model serving game
blog.cloudflare.com/workers-ai-lar…
English
@ashishbdatta @ashishbdatta quick comparison:
Black bar sections show incremental percentage for AWQ over the bf16 model.
Not much difference
English
Ashish retweetledi
3 fresh updates to Workers AI:
⚡️ 2x–4x faster inference with speculative decoding & prefix caching
⚡️ Batch inferencing for handling large request volumes seamlessly
⚡️ More LoRA options for customized models
blog.cloudflare.com/workers-ai-imp…
English
Please route everything I broke to ->
michelle@michellechen
some workers ai devweek changes are rolling out... starting early and going slow. if you see anything off with your model outputs, hit my DMs
English
Ashish retweetledi
Llama 4 Scout 17B Instruct is now available on Workers AI: use this multimodal, Mixture of Experts AI model on Cloudflare's serverless AI platform to build next-gen AI applications.
blog.cloudflare.com/meta-llama-4-i…
English
Ashish retweetledi
saturday scout surprise
blog.cloudflare.com/meta-llama-4-i…
English
