Shritesh Bhattarai Profile
Shritesh Bhattarai

@shritesh

Followers
885
Following
15K
Media
7
Statuses
487

CTO @dylibso // Making software extensible with AI

San Francisco, CA
Joined April 2009
Don't wanna be here? Send us removal request.
@shritesh
Shritesh Bhattarai
6 days
I do 99% of my dev work inside Conductor now. Go work with Charlie and make it 100%.
@charlieholtz
Charlie Holtz
7 days
come work with us at @conductor_build! we're growing fast and have big plans. interfaces to orchestrate AIs are already changing how we work and we think we have a real shot at building the best one. if you're up to moving to San Francisco send me a DM or apply below! DAU
1
0
9
@steipete
Peter Steinberger šŸ¦ž
8 days
AI lets you move fast, but getting the details right is still hard. So many edge cases. That's the difference that makes sth feel like slop or polished.
45
21
490
@josevalim
JosƩ Valim
17 days
Today we celebrate 15 years since Elixir's first commit. To mark the occasion, we shipped an early release candidate for Elixir v1.20 which has TYPE INFERENCE FOR ALL CONSTRUCTS.šŸŽ‰ In the latest article, I break down the next ~15 weeks and the challenges for the next ~15 months.
25
110
569
@andy_pavlo
Andy Pavlo (@andypavlo.bsky.social)
22 days
Here is my latest article on the world of databases: https://t.co/hngkeN0dlN All the hot topics from the last year: • More Postgres action! • MCP for everyone! • MongoDB gets litigious with FerretDB! • File formats! • Market movements! • The richest person in the world!
Tweet card summary image
cs.cmu.edu
The world tried to kill Andy off but he had to stay alive to to talk about what happened with databases in 2025.
28
165
960
@davesnx
David Sancho
25 days
I built my own JavaScript runtime in OCaml Mostly, I wanted to see how far OCaml could go against Bun and Node, and also OCaml vs zig in similar APIs. So I built `sirocco`. it uses JavaScriptCore (same engine as Bun and Safari). Event loop with libuv and it's written in
25
55
404
@nilslice
steve.wasm
29 days
šŸ’Æseeing this play out: 1. set up the model for successful outputs 2. simplify the runtime semantics, make powerful primitives hard to misuse 3. have deterministic verification providing feedback to model on its outputs 4. design for output composability and integration
@martin_casado
martin_casado
1 month
When it comes down to it, with AI, much of software reduces to dealing with fundamental trade offs of runtime semantics. This is arguably what the bulk of distributed systems and database development is already. It'll be interesting to see AI push the state of the art here.
1
1
2
@notnotrishi
Rishi
1 month
After looking at patterns of failure modes during generation for 4 days...surprisingly most of them were not related to how models 'reason' -- feel like more of 50y old (CS) compiler inspired techniques/constraints could be applied to solve many issues in LLM inference (decoding)
1
1
1
@threepointone
sunil pai
1 month
all these ā€œwebsitesā€ suck and I’m tired of pretending like they don’t the web could do with a lot more ambition and fewer apologists
16
8
241
@shritesh
Shritesh Bhattarai
1 month
If not compiler problem, why compiler problem shaped?
0
1
2
@wyeworks
wyeworks
2 months
At the San Francisco Elixir Meetup watching @josevalim talking about programming language evolution in the #AI era #MyElixirStatus
0
1
4
@jeffreyhuber
Jeff Huber
2 months
I'm hosting an event on Dec 2 on the topic of Agentic Search. Learn where search is going next including a preview of our next research report. https://t.co/ReFctIq04v
Tweet card summary image
luma.com
For anyone that has used Claude Code, Codex, or Deep Research - the direction is clear - search is becoming fully agentic. Agentic search enables us to tackle…
3
2
30
@HelmPack
Helm
3 months
ā€¼ļøHelm v4.0.0 is now officially live! https://t.co/31KwAS8lwkā€¼ļø New updates include: šŸŽ‰ plugin system that supports WASM based plugins šŸŽ‰ Server side apply support šŸŽ‰ & LOTS more Go check it out! & if you’re around at #KubeCon, come chat about it at the Helm booth!
Tweet card summary image
github.com
The Helm Team is proud to announce the first stable release of Helm 4. New Features Helm 4 has numerous new features, but a few deserve highlighting here: Redesigned plugin system that supports We...
4
76
253
@extism
extism
2 months
Much love to the @CloudNativeFdn Helm team using Extism to power v4's new plugin system. https://t.co/9apK0yZEdd šŸŽ‰ Congrats on the new release!
2
4
7
@neobrain_
Tony Wasserka
3 months
Cat's out of the bag: FEX, an emulator I've been working on these past 4 years, is going to be part of the Steam Frame to bring x86 Windows games to an ARM Linux VR headset! 😊 https://t.co/R8OeosyySd
62
154
2K
@FelixCLC_
@fclc
3 months
LETS FREAKING GOOOOOOOOOOOO FP16 ON X86 DESKTOP!!!
@InstLatX64
InstLatX64
3 months
#AMD #Zen6 znver6 ISA: - #AVX512_BMM (CPUID.80000021.EAX[23], VBMACOR16x16x16, VBMACXOR16x16x16, VBITREV) - #AVX512_FP16 - #AVX_NE_CONVERT - #AVX_IFMA - #AVX_VNNI_INT8 Source: https://t.co/hmQSDXlHpD
11
11
203
@renerocksai
renerocksai⚔
3 months
Phew! "The Writeup" about https://t.co/sKNXt2KE5m, my #TigerBeetle ticketing demo, is done! Read about the whole journey from just an idea to 977 ops/s here: https://t.co/gxVjEqFUm1
2
15
51
@chris_mccord
Chris McCord
3 months
Here's a little writeup about how `mix fly_deploy.hot` works. It's just basic fly primitives + Erlang stdlib. Erlang just has all these ridiculous features sitting there https://t.co/EpRLCS8QN9
Tweet card summary image
community.fly.io
Today I published FlyDeploy, an Elixir package providing mix fly_deploy.hot for hot code upgrades of your Elixir and Phoenix applications. This allows you to deploy changes to your running apps...
1
14
110