Atul Shah

Entrepreneur. Tech Enthusiast. Angel Investor. Believer in Underdogs. India Optimist. Opinions expressed are mine.

[Atul Shah @atulshah]

@adithya_s_k Does it also support recognition of handwritten documents ?

[Atul Shah @atulshah]

@karpathy @karpathy Well written review!
A couple things -
1. How good is the tech for crowded street navigation where traffic rules may not be followed?
2. How near or far are the Chinese car competitors?

[Atul Shah @atulshah]

@p_gururani Agree that SLMs need more adoption in Enterprises...
https://t.co/xzr9f4uUuq

[Atul Shah @atulshah]

@DeepikaBhardwaj Sad but true.... people don't value good advice.

[Catalin @catalinmpit]

Whaaaat? Is WindSurf shutting down?

[Atul Shah @atulshah]

I'm on the waitlist for @perplexity_ai's new agentic browser, Comet: https://t.co/iRqvhZWpzf

[Atul Shah @atulshah]

My @boltdotnew hackathon project, Rootd, is live!
It's our quiet rebellion against digital noise. A simple app to build real character, not a social profile.
One prompt a day. No scores, no streaks.
Demo: https://t.co/MGTVrIktNS
Website: https://t.co/wsPpNYw9K2
#bolthackathon

[Atul Shah @atulshah]

"Built at the world's largest hackathon! by @boltdotnew
Rootd: character-building through daily prompts of Courage, Compassion & Cheerfulness https://t.co/wsPpNYw9K2
Check this Rootd Philosophy page for more details

[Andrej Karpathy @karpathy]

+1 for "context engineering" over "prompt engineering".

People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window with just the right information for the next step. Science because doing this right involves task descriptions and explanations, few shot examples, RAG, related (possibly multimodal) data, tools, state and history, compacting... Too little or of the wrong form and the LLM doesn't have the right context for optimal performance. Too much or too irrelevant and the LLM costs might go up and performance might come down. Doing this well is highly non-trivial. And art because of the guiding intuition around LLM psychology of people spirits.

On top of context engineering itself, an LLM app has to:
- break up problems just right into control flows
- pack the context windows just right
- dispatch calls to LLMs of the right kind and capability
- handle generation-verification UIUX flows
- a lot more - guardrails, security, evals, parallelism, prefetching, ...

So context engineering is just one small piece of an emerging thick layer of non-trivial software that coordinates individual LLM calls (and a lot more) into full LLM apps. The term "ChatGPT wrapper" is tired and really, really wrong.

[Andrej Karpathy @karpathy]

It's 2025 and most content is still written for humans instead of LLMs. 99.9% of attention is about to be LLM attention, not human attention.

E.g. 99% of libraries still have docs that basically render to some pretty .html static pages assuming a human will click through them. In 2025 the docs should be a single your_project.md text file that is intended to go into the context window of an LLM.

Repeat for everything.

[Atul Shah @atulshah]

@grok @grok on what dataset have you been trained on. How latest is it?

[Atul Shah @atulshah]

@elonmusk @grok When is Grok coming to Android

[Atul Shah @atulshah]

I had written about "vibe-coding" in my LinkedIn post https://t.co/bSioCkO4Ak
John Rush's thread dives into 36 AI coding agents & IDEs like Cursor (Claude 3.7), Windsurf, & Replit, boosting dev productivity. A must-read on the future of coding with AI! @johnrushx #AICoding ".

[John Rush @johnrushx]

I've tried all (36) AI Coding Agents & IDEs 😵‍💫

[CreateXyz, Cursor, Softgen, Windsurf, Wrapifai, Copilot, Lovable, Bolt, v0, Replit, MarsX, AmazonQ, Pear, Devin, Github Spark, IDX, Webdraw, Claude 3.7 Sonnet & more]

The most complete list ever made:

[Atul Shah @atulshah]

Extremely valuable insights by @kmr_dilip for #HealthTech #Founders .

[Dilip Kumar @kmr_dilip]

Have met over 600+ healthcare startups this year at Rainmatter. These are my general observations of the sector in 2024 so far. An investor & health enthusiast perspective.

Remember, these are observations and not opinions. If you’re a builder, I'm always keen to exchange notes.

1) Health tech needs maturing – The novelty phase of tech meets health is over and the focus should be on driving outcomes to the user.

2) Preventive health is gaining traction – More startups are focused on preventive solutions.

3) Nutrition personalisation is heating up – Companies are going beyond generic consulting and leveraging DNA, microbiome, or other personal data for nutrition recommendations.

4) UPI for health not yet in India- ABDM use cases haven't seen much adoption.

5) Wearables are no longer the star – Smartwatches and rings are still strong but are considered table stakes. Consumers want something that tells “why” and not limited to “what”.

6) Corporate wellness is an overrated market – With rising mental health issues, stress, and burnout in the workforce, B2B wellness solutions ideally should show promise. But there isn’t adoption from corporates.

7) The “biohacking” trend is niche but growing –longevity startups are speculative but hot. Increasing interest in biohacking protocols (sleep optimization, anti ageing etc.). But the market is niche and often hard to validate scientifically unless clinical protocols are done.

8) Consumer trust in health supplements is dwindling – The supplement industry lacks regulation and transparency and this has eroded trust amongst consumers. Brands are focusing on science-backed products.

9) AI in healthcare – Many are bullish on AI’s potential to revolutionise diagnostics, patient care, and drug discovery, but deployment in clinical settings is slow and capital-intensive.

10) Mental health is booming – Digital mental health startups, especially those offering therapy, coaching, or CBT solutions are seeing adoptions but differentiation is becoming a challenge in a crowded market.

11) Community-driven fitness – Social, community-oriented fitness solutions still have potential if done right especially those blending real-world interactions with digital offerings.

12) CAC is a challenge for D2C health brands – Brands are facing rising customer acquisition costs often with no sustainable GTM.

13) Hybrid health businesses are scaling better – Companies combining digital and physical services have better scope of scaling fast.

14) Telehealth plateaued – The initial excitement around telehealth has subsided for both consumers and investors.

15) Data privacy concerns – As health tech companies gather more sensitive data, concerns about privacy and security will become critical.

16) Health marketplaces have no moat – Platforms connecting consumers with healthcare providers or wellness services are overrated with no clear moat to grow.

17) Clinical validation is key – Companies that can prove clinical efficacy or show robust evidence for their solutions are attracting more interest both from consumers and investors.

18) Wellness is mainstream, but consumers are cynical – Consumers are more wellness-conscious than ever, but they are also more sceptical of unproven claims, forcing brands to lead with science and transparency.

19) Building trust is harder than ever – With consumer scepticism at an all-time high, companies need to prioritise transparency, scientific validation, and clear communication to gain consumer trust, especially in nutrition.

20) Retention over acquisition – CAC have skyrocketed and focus should be on retention. Building loyal, long-term customers by driving outcomes will be key

21) Offline isn’t dead – Post pandemic people are realising that in-person experiences matter more. The winning strategy seems to be a hybrid model that leverages both digital convenience and real-world engagement.

22) Getting clinical validation is expensive but essential – Especially those offering treatments or diagnostics, obtaining clinical trials or peer-reviewed studies is critical for credibility. But the cost and complexity can be daunting.

23) Customer education is a feature – Whether in fitness or nutrition, startups can’t just offer products; they need to educate consumers on why their solution matters. Content marketing and community-building should be the main features.

If you’re a builder, I'm always keen to exchange notes.
We're always investing and cheering for founders at @Rainmatterin.

[Atul Shah @atulshah]

Dr. Mashelkar talking about Heath innovation .. Amazing..Pole-Vaulting instead of Leap-Frogging @rameshmashelkar Health Technology Group BHARAT in Pune by @HTGBharat
#HTGBharat #HealthTechPune

[Atul Shah @atulshah]

Hello @mngl_in . Its been 6 months and still no connectivity even after taking the deposit of 6500/- more than 2 yrs back. 6months back in your tweet you had said that 15-20 days work will get completed

[Maharashtra Natural Gas Ltd. @mngl_in]

@atulshah The above activity is planned in the next 15-20 days for needful further to complete the connectivity and provide the pending connections in your Society.

[Ritesh Jayswal @rjayswal]

@atulshah Kumar Sanu flashback time

[Atul Shah @atulshah]

@airtelindia @Airtel_Presence
I am an Airtel Black customer. #Airtel Xtream fiber down for last 4 days. Chatbot interaction on App that keeps on saying Technical issues without being able to raise a complaint. There is no support phone or person available .

[Atul Shah @atulshah]

@mngl_in connection deposit has been taken from residents in our society (ICS Colony , Pune) for more than a year. Yet MGNL connection has not been provided. Area service mgr Omkar not responding . Marketing manager who came to collect the docs & cheque says contact area mgr