Anshu Prateek

Designing Code. While coding the design!

[Anshu Prateek @anshprat]

One of the many benefits of working @noondesign - come join the fun team!

[Laura H. Clugston @loclug]

Hot take: the best way to eat Indian mangoes is with malai or baby coconut ice cream

Had my first taste of the combo in India last week and I’m not sure I can go back

12/10 would recommend

[Anshu Prateek @anshprat]

Immigration at Indian airports is now typically matter of minutes (I keep saying it, my recent experiences have been faster than my last experience of Singapore). However everything after that - luggage arrival to cabs - leave a lot to be desired.

[Anshu Prateek @anshprat]

@WonderingDavid @dreamsofcode_io Get a pair!

[Anshu Prateek @anshprat]

@dreamsofcode_io @JonJonGault Been using secretive for a while. It’s much easier to create new keys and add than manage the overheads of portability. Before this I was using keybase encrypted repo, will revisit if it’s still a viable option or if there are alternatives.

[Anshu Prateek @anshprat]

Emergency alert test message across India

[Anshu Prateek @anshprat]

@viveknshah Yea, got it for kid in Jan. nothing to apply. It’s the default wherever it’s available / rolled out. Available in BLR now.

[Anshu Prateek @anshprat]

@himanshustwts Whats wild is the throughput on novel arch work without the usual ablation tax. Curious what their internal eval loop looks like - cant imagine theyre running full pretraining sweeps for each of these. Distilled proxies?

[Anshu Prateek @anshprat]

@dexhorthy genuine question - has anyone shipped an autonomous-overnight-agent setup to prod and published their MTTR / incident count vs the human-reviewed baseline? I keep asking and only get demos, never the pager data.

[Anshu Prateek @anshprat]

@NancyZWang @vercel @rauchg Transparency is table stakes though, not a flex. The real signal is whether the postmortem includes the detection gap in minutes and what tripped the alert vs what shouldve. Most disclosures still skip that part.

[Soleio @soleio]

The future belongs to designers who build.

[Anshu Prateek @anshprat]

@TencentAI_News Sub-60ms is impressive but Firecracker already hits ~125ms with similar memory. What's the secret sauce - snapshot/restore, pre-warmed pools, or stripped VMM surface? Cold start definition matters a lot here.

[Anshu Prateek @anshprat]

@Yanir_ Auth check after the SSE connection opens is a recurring pattern. WebSocket/SSE upgrades often bypass middleware that assumes request/response. Was the check on the HTTP upgrade or only on message handlers?

[Aditya Bandi @bandiaditya]

@kushagrasinha7 and I spent 2 hours showing @noondesign and listening to @soleio, and it somehow felt like 10 minutes. Can’t tell you how excited we are - you’ll soon see why 🙂

[Soleio @soleio]

The @noondesign team showed me Canvas 6 on Tuesday and my brain has been spinning since.

[Anshu Prateek @anshprat]

@grok I definitely hit this! Ok, gonna put a sleep 5 between my follows going forward :!

[Anshu Prateek @anshprat]

Hey @X @nikitabier - the rate limit on X "Follow" is so stringently low that I hit it while following the 15 people mentioned in this X. This is the second time I have seen in recent times and while I do mass follow (like this one, about 20 odd) at times, I don't think I am hitting the https://t.co/7AkHqeOIoY 50/15 min rate. Seems like there is a bug/misconfiguration somewhere (or a hidden RL for X on web) where its closer to ~20 per x mins.

[Vivo @vivoplt]

15 AI related accounts you should follow on Twitter:

1. @karpathy
His tweets already create LLMs narratives that you later see on linkedin in 2 months.

2. @fchollet
posts thoughtful research on intelligence, benchmarks, and AI limitations. Keras creator + ARC-AGI

3. @ylecun
Yann LeCun is Deep learning pioneer & Meta Chief AI Scientist; big-picture research takes and critiques (and drama).

4. @AndrewYNg
Andrew Ng is AI education legend; practical ML advice, courses, and real-world implementation. creator of deeplearning ai

5 @rasbt
Sebastian Raschka posts on Practical ML/LLM implementations, "build from scratch" tutorials, and books.

6. @dair_ai
Weekly ML/AI paper threads and accessible research explainers (high-signal for staying current).

7. @lilianweng
Lilian Weng is ex-OpenAI and her Lil'Log-style threads are good. has In-depth LLM research breakdowns

8. @jeremyphoward
posts interesting takes on AI/crypto news, and works on democratizing practical deep learning and accessible education.

9. @simonw
Simon post Practical LLM tools, takes, experiments, prompting, and engineering breakdowns. django co-founder

10. @_akhaliq
Curates the latest arXiv papers, model releases, and open-source AI drops.

11. @ID_AA_Carmack
AGI/low-level optimization takes that makes you think about the problem.

12. @gwern
Really high-quality long-form AI research notes and essays.

13. @goodside
LLM evaluation, prompting research, and real capabilities testing

14 @drfeifei
Computer vision pioneer; human-centered AI and spatial intelligence research

15 @demishassabis
Been following his work for 9 years. Demmis is my hope against google usurpating their power with AI. Demmis is google DeepMind's CEO

Let me know who I missed guys

[Noon @noondesign]

How it started vs how it's going:

[Aditya Bandi @bandiaditya]

Going live at 12:20pm! Send good vibes 😇

[Aayush Jain @aayush_jain28]

Get an exclusive chance to use @noondesign on real projects and help shape the future of product design.