anshuman (@athletic_coder) profile picture

anshuman

@athletic_coder

mostly ml here;

heyyanshuman.com
Joined February 2020
  • Tweets10.5K
  • Following1.1K
  • Followers20.6K
1.2KMedia
Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)Photo by anshuman (@athletic_coder)
anshuman retweeted
anshuman (@athletic_coder) profile picture
anshuman
@athletic_coderNov 5
this aged really well
Image from anshuman (@athletic_coder)
8
72
3
1,913
45,891
this aged really well
Image from anshuman (@athletic_coder)
8
72
3
1,913
45,891
i think about this a lot
swyx 🔜 aie.codes (@swyx) profile pictureswyx 🔜 aie.codes
@swyxMar 9
btw if you can build a category leader open source project in ai engineering right now the market acquihire rate is ~$10-$100m per ai engineer. you do not need to figure out a business model, you do not need GTM, you do not need funding. just build things clankers want.
3
1,609
anshuman retweeted
anshuman (@athletic_coder) profile picture
anshuman
@athletic_coderNov 3
INFERENCE is all you need🔥 working on this detailed blog on KV Cache Engineering for Agentic Inference starting from basics we are gonna explore in DEPTH the role KV Cache and tradeoffs associated with it in production settings stay tuned.
Image from anshuman (@athletic_coder)
Image from anshuman (@athletic_coder)
10
1
18
1,395
high signal, no bs channel must must must must watxh
swyx 🔜 aie.codes (@swyx) profile pictureswyx 🔜 aie.codes
@swyxNov 2
after 3 years of building from nothing, we are now serving 2M monthly unique engineers on aie youtube. (76% not subscribed lol. idk if it matters but pls subscribe! we need to keep up momentum to cross 1M subs roughly around aie.codes it looks like)
Image from swyx 🔜 aie.codes (@swyx)
2
942
INFERENCE is all you need🔥 working on this detailed blog on KV Cache Engineering for Agentic Inference starting from basics we are gonna explore in DEPTH the role KV Cache and tradeoffs associated with it in production settings stay tuned.
Image from anshuman (@athletic_coder)
Image from anshuman (@athletic_coder)
10
1
18
1,395
one of those bad sleepless nights
2
777
hyperbolic is onto something. first neocloud, afaik to hire a analyst, quant etc
Jasper (@zjasper) profile pictureJasper
@zjasperNov 2
Very excited to welcome @blu_putnam, former Chief Economist of CME Group, to Hyperbolic as our Senior Research Analyst. As GPU compute becomes a tradable market, his perspective will be invaluable. His first piece: why today's indices can't agree on what a B200 hour costs.
1
783
anshuman retweeted
anshuman (@athletic_coder) profile picture
anshuman
@athletic_coderNov 2
it has begun
Image from anshuman (@athletic_coder)
4
4
1
130
10,344
it has begun
Image from anshuman (@athletic_coder)
4
4
1
130
10,344
smol career update: got promoted to Applied Scientist 2
Image from anshuman (@athletic_coder)
25
8
560
12,472
smol career update: got promoted to Applied Scientist 2
Image from anshuman (@athletic_coder)
25
8
560
12,472
I need VC funding. C=Compute
2
9
1,300
never leaving this app
Image from anshuman (@athletic_coder)
1
5
1,159
never leaving this app
Image from anshuman (@athletic_coder)
1
5
1,159
banger
Mehul Mohan (@mehulmpt) profile pictureMehul Mohan
@mehulmptOct 23
can't wait for people to share guides on how to host 3 trillion kimi k3.1 pro model on their laptop getting 1000s/t speed
1
914
RSI is the new AGI
3
871
I put together a comprehensive resource list for learning model post-training and building useful applications. now i'll be sharing the resources one by one. this is my post-training resources series. save this as your starting point. links in thread 🧵
Image from anshuman (@athletic_coder)
1
6
10
1,269
very curious if they used deepseekharness for automationbench maxxing
DeepSeek (@deepseek_ai) profile pictureDeepSeek
@deepseek_aiOct 17
🚀 DeepSeek-V4.1-Flash-1017 is here! Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API. 🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model. 1/4
Image from DeepSeek (@deepseek_ai)
Image from DeepSeek (@deepseek_ai)
6
1,857
$0.75 / 1000 searches ???????????? exa is cooked.
DeepSeek (@deepseek_ai) profile pictureDeepSeek
@deepseek_aiOct 17
Replying to @deepseek_ai
🔍 Search & Fetch APIs are now open to everyone! Powered by our in-house search engine, serving hundreds of millions of requests daily worldwide. 💰 Each endpoint: $0.75 per 1,000 calls. 📖 Docs: api-docs.deepseek.com/guides… 📌 Search & Fetch calls via server-side tools in DeepSeek Harness, Claude Code, and Codex are currently free for DeepSeek API users. Billing starts Oct 31, 2026, at 16:00 UTC. 3/4
Search & Fetch | DeepSeek API Docs

Search & Fetch | DeepSeek API Docs

The Search API and the Fetch API give agents and RAG applications access to live web content:

api-docs.deepseek.com
53
85
3
3,908
281,431
Load more