Vedant Singhal

👋
Curious mind exploring tech, design & ideas that shape the future.
Engineering Student @iitdelhi
🐦‍🔥

All views are mine and personal.

[Vedant Singhal @its_vedants]

@Siddhartha37648 Totally agree... It seems more of a pay to win due to limited compute and access to AI Agents for iterating.

[Vedant Singhal @its_vedants]

@altitude_jaden Are you also hiring interns?

[Vedant Singhal @its_vedants]

@Ashneer_Grover If you reduce freebies, UPI running costs can already be covered very easily anyways. Freebies are anyways for controlling vote banks which should not be happening.

[Vedant Singhal @its_vedants]

@m_saharia Is the code open source yet?

[Vedant Singhal @its_vedants]

@_Creation22 Hey there!

[Sam Altman @sama]

You can skip all the parties, all the conferences, all the press, all the tweets. Build a great product and get users and win.

[Vedant Singhal @its_vedants]

Seems like a very promising model and definitely going to lead the way for some time now. The improved forecast accuracy, combined with the ability to leverage live satellite data, could make it significantly more useful for real-time and short-term weather prediction.

[Google AI @GoogleAI]

Introducing WeatherNext 3️⃣— our most advanced global weather AI model yet from @GoogleDeepmind and @GoogleResearch

With prediction capabilities that are up to 5x sharper than WeatherNext 2, the model generates a forecast with high spatial resolution in order to catch fast-evolving rainstorms, map local temperature shifts, and even help wind farms predict their power output.
So, how does it do that?

While traditional weather models rely on massive, physics-based supercomputer simulations that can carry a 6-hour forecast lag, WeatherNext 3 leverages live geostationary satellite observations as inputs and trains directly on real-world surface and atmospheric observations.

By pulling this raw satellite data, it’s able to update the global forecast every single hour. And because weather develops at lightning speed, these quick, detailed insights can help bring more localized forecasting to billions of people and local businesses, especially in regions that are historically underserved due to the high costs of traditional weather forecasting models.

[Vedant Singhal @its_vedants]

@prajdabre Just curious... How does one get into the internal testers of these ai models?

[Vedant Singhal @its_vedants]

@solitrix02 @Shrii_6 How do you get the clients?

[Vedant Singhal @its_vedants]

@DrDatta_AIIMS The media has been interacting with a set of people only and the same faces are on every channel.

[Vedant Singhal @its_vedants]

@MeteKarsli0glu @eric_nilsson_ So where will rain happen then? Can you please tell a bit more about what your drones do?

[Vedant Singhal @its_vedants]

@cb_doge @Starlink Can you please tell one thing:
"Rural subscription density was 48.31 per 100 people, compared with 126.80 in urban India."
How is the number even greater than 100?

[Vedant Singhal @its_vedants]

@SarvamAI Excited to see you guys at IIT Delhi!

[Andrew Ng @AndrewYNg]

New: A map of the most important skills in AI Engineering. https://t.co/VVkn1Dqp1N

[Vedant Singhal @its_vedants]

Huge Congratulations & Really Cool!!!

[Skyroot Aerospace @SkyrootA]

ORBIT ACHIEVED. 🚀

Vikram-1 Test Flight-1 has reached orbit. India's first privately developed orbital rocket has completed its final burn and injected its payloads into a ~450 km orbit, making India the third country in the world with private orbital launch capability.

History is made. 🇮🇳

#Vikram1 #JourneyToOrbit #SkyrootAerospace

[Vedant Singhal @its_vedants]

Then I asked it to redesign the pipeline for lower latency. It didn’t suggest optimizations—it reasoned about caching, concurrency, state, and failure recovery together. That’s when GPT-5.6 stopped feeling like autocomplete and started feeling like an engineer reviewing a design.

[Vedant Singhal @its_vedants]

3/
What impressed me wasn’t file-level explanations—it understood why the system was built that way.
It saw why expensive AI video was separated from cheap FFmpeg overlays, so changing a name or date never requires regenerating the entire video.
That’s architectural reasoning.

[Vedant Singhal @its_vedants]

@codewithsushi The irony is unreal—you criticize others on Instagram for chasing engagement, and then turn around asking people to comment just to boost your own reach. Wow.

[Vedant Singhal @its_vedants]

@GoogleAIStudio All of this is fine but when are you giving Gemini 3.5 Pro?

[Vedant Singhal @its_vedants]

@cneuralnetwork I got both DP questions in coding round