Gitanshu Munjal

crops, statistics, plant breeding, poker, stocks, predicshun! #UCDavis PhD. Building a kind of digital breeder’s eye 🌽👁️ @ EWS. Opinions are entirely mine.

[Sparsh Agarwal @sparshselim]

The latest issue written by @jainhiya_ delves into the history of Indian Statistics.

The story of Mahalanobis is peppered with tales of creative ingenuity, scrappy technological jugaad (like India’s first analog computer), & the coming together of a newly independent nation’s smartest scientists.

[Hebihime @onepiecei5drug]

What I love about Luffy is how simple and straightforward he is. He doesn’t dwell on emotional conflicts or overthink everything. He says what’s on his mind and keeps moving forward. I really aspire to have that kind of mindset…

[Gitanshu Munjal @grmunjal]

@jxnlco caffeinate -dims

[Gitanshu Munjal @grmunjal]

Fun to ask it plant breeding questions

[Nick Levine @status_effects]

New work with @AlecRad and @DavidDuvenaud:

Have you ever dreamed of talking to someone from the past? Introducing talkie, a 13B model trained only on pre-1931 text.

Vintage models should help us to understand how LMs generalize (e.g., can we teach talkie to code?). Thread:

[Bo Wang @BoWang87]

New @Nature paper today : Sony's Ace robot beats 3 of 5 elite table tennis players. Loses to professionals.

Human players win points with faster-than-average shots (p<0.001 between won vs returned). Ace wins with ordinary shots. Same speed and spin profile whether it wins or loses the rally (p=0.88).

It's playing a completely different sport than the humans are.

Trained entirely in simulation. Zero sim-to-real tricks beyond good physics modeling and asymmetric actor-critic (critic sees ground truth, actor sees noisy sensors).

Best part — after watching a point, 1992 Olympian Kinjiro Nakamura said:

"I didn't think it was possible. But the fact that it was possible... means that there is a possibility that a human could do it too."

Code: https://t.co/8ZZEh3BeVm

Paper: https://t.co/NhFPzYVOyi

[Gitanshu Munjal @grmunjal]

@reach_vb @dejavucoder Shooting my shot, sent a DM! 🥹

[Sparsh Agarwal @sparshselim]

the craft of kanchipuram, the intelligence of diffusion models, the creativity of @AltCarbonIndia's design team, the magic of @AlterMagIndia's latest piece

a feast for the mind, a carnival for the eyes

https://t.co/UTL7e5ipXh

[Async @asynchronous_x]

Opus 4.6 burning 250k tokens and dimming the lights of a small village just to adjust the tailwind padding on my div wrong

[Gitanshu Munjal @grmunjal]

=AI() in a google spreadsheet is the stuff nightmares are made of

[Uncle Bob Martin @unclebobmartin]

With one agent, I used to wait for Claude. With two agents I still waited for Claude, but not as long. With three agents Claude is waiting for me. I am the bottleneck. And the bottleneck is all planning.

[Gitanshu Munjal @grmunjal]

@levelsio @AbzGupta @5aitec @UltrahumanHQ

[shyamal @shyamalanadkat]

very exciting to launch ChatGPT Go, a new subscription tier in India at ₹399/month that gives users 10x higher message limits, 10x more image generations, 10x more file uploads, and 2x longer memory compared with free tier. and you can now pay with UPI!

🇮🇳 वन्दे मातरम्

[Ryo Lu @ryolu_]

There is no secret sauce to making something great.

It just takes a group of believers who care, who are unafraid to build and try crazy ideas.

https://t.co/JwnNQCGS2P

[Gitanshu Munjal @grmunjal]

@shyamalanadkat *for knowledge workers in metropolitan cities

The big upgrades to the free tier are my favourite from this release. The 3.5 to 4 type of feeling I had, I see a lot of people around me having recently.

[Gitanshu Munjal @grmunjal]

But also the presentation🤌 They don’t write “harnessing AI for revolutionizing language” reviews; they drop web demos, containerized code, and leaderboards. In contrast we get buzzword reviews or throwaway model for trait x papers with shallow architecture recaps and glue code

[Gitanshu Munjal @grmunjal]

Wild contrast in reading new breeding / applied genetics papers vs ai papers. More than a decade later the first mostly still GWAShing for trait x in crop y. Second is so much more fun because there is novelty and curiosity in abundance not in just a small percent of papers.

[Gitanshu Munjal @grmunjal]

@pastramimachine Yes!!

[Gitanshu Munjal @grmunjal]

Absolutely HATE what any spreadsheet software enables people to do. Most spreadsheets are people thinking they are doing math or record keeping but actually creative writing.

[Gitanshu Munjal @grmunjal]

🙌

[Dilip Kumar @kmr_dilip]

India was once called a nation of call centers and it sounded like an insult. But quietly, it trained millions of young Indians to speak fluent English and dream beyond their hometowns. That was phase one.
Then we became the IT back office of the world. And it created a generation of engineers who could write code, manage complex systems, and deliver at global scale. That was phase two.

Today, we’re in phase three. Indian startups are solving for India and it looks chaotic from the outside. These are signs of a country solving its own problems first, before solving the world’s.

Sure, we should go more deep tech and we will i think. But first we need more patient capital, more research led universities & more talent density. Deep tech ecosystem can't be built in a hackathon. It will be built by million engineers who get obsessed in solving billion daily problems.