Selvakumar R

Building the future in the AI wilderness.
IIM BANGALORE grad.

[Selvakumar R @R_Selva_kumar]

Once upon a time , internet used to be fun

[Selvakumar R @R_Selva_kumar]

Intelligence cannot be boxed because of its own virtue. Regards to closed vs open weight models, the delta will will flow distilled as weights or as skills ( cleaner way) from frontier models.
For a restaurant, not just chef, the recipe matters and the ambience matters.

[Selvakumar R @R_Selva_kumar]

Network effects is one moat , but thats getting wasted in doomscrolling.
The infinite intelligence has time travelled from future and gave us doomscrolling, while we scroll, its growing exponentially

[Selvakumar R @R_Selva_kumar]

Intelligence cannot be boxed because of its own virtue. Regards to closed vs open weight models, the delta will will flow distilled as weights or as skills ( cleaner way) from frontier models.
For a restaurant, not just chef, the recipe matters and the ambience matters.

[Selvakumar R @R_Selva_kumar]

The big shift is the infinite intelligence is making the human defined moats look as rounding error when compared to the infinite intelligence

[Selvakumar R @R_Selva_kumar]

@8teAPi Claude Tag is ambient and more natural way to build the Ontology and will perform better than human FDEs. But organisations will self select either of the approaches. Similar to Slack vs teams , outlook vs gmail

[Selvakumar R @R_Selva_kumar]

@kunalb11 Congratulations Kunal

[Cristian @cristiannmillo]

Esto es espectacular como se movió Marruecos defensivamente. Esto se entrena. No es aleatorio. Tremendo bloque corto defensivo. Imposible de entrar sin alguna magia o pase filtrado con extrema exactitud. Por eso Brasil se la pasó lateralizando.

[Selvakumar R @R_Selva_kumar]

@nikesharora Similar to " Intel Inside" in a computer era. It will be " intelligence inside" in AI era. Apps layer it is again

[Selvakumar R @R_Selva_kumar]

@KP24 Interesting and exciting times. T20 version 2.0

[Beff (e/acc) @beffjezos]

The rarest object type in the universe isn't black holes. It's us. Conscious matter. The flame of life.

We have a duty to expand it in scope and scale in order to preserve it.

[Selvakumar R @R_Selva_kumar]

@TheAhmadOsman 😂

[Bryan Johnson @bryan_johnson]

If Hantavirus mutated into a global threat, it would unleash AI + biotech unlike anything we've ever seen.

> genome sequenced and public in 4 hours
> AlphaFold maps every protein target
> AI screens 10,000 drugs in 24 hrs
> 50 vaccine candidates designed simultaneously
> AI designed antibodies in days
> risk of death computed instantly
> decentralized trials launch globally
> enroll from home
> 20 countries manufacturing at once
> first doses in three weeks
> real-time dose characterization
> your genome + biomarkers determine your protocol
> variant map updates every hour

No one would wait for governments.

[Andrej Karpathy @karpathy]

Fireside chat at Sequoia Ascent 2026 from a ~week ago. Some highlights:

The first theme I tried to push on is that LLMs are about a lot more than just speeding up what existed before (e.g. coding). Three examples of new horizons:

1. menugen: an app that can be fully engulfed by LLMs, with no classical code needed: input an image, output an image and an LLM can natively do the thing.
2. install .md skills instead of install .sh scripts. Why create a complex Software 1.0 bash script for e.g. installing a piece of software if you can write the installation out in words and say "just show this to your LLM". The LLM is an advanced interpreter of English and can intelligently target installation to your setup, debug everything inline, etc.
3. LLM knowledge bases as an example of something that was *impossible* with classical code because it's computation over unstructured data (knowledge) from arbitrary sources and in arbitrary formats, including simply text articles etc.

I pushed on these because in every new paradigm change, the obvious things are always in the realm of speeding up or somehow improving what existed, but here we have examples of functionality that either suddenly perhaps shouldn't even exist (1,2), or was fundamentally not possible before (3).

The second (ongoing) theme is trying to explain the pattern of jaggedness in LLMs. How it can be true that a single artifact will simultaneously 1) coherently refactor a 100,000-line code base *and* 2) tell you to walk to the car wash to wash your car. I previously wrote about the source of this as having to do with verifiability of a domain, here I expand on this as having to also do with economics because revenue/TAM dictates what the frontier labs choose to package into training data distributions during RL. You're either in the data distribution (on the rails of the RL circuits) and flying or you're off-roading in the jungle with a machete, in relative terms. Still not 100% satisfied with this, but it's an ongoing struggle to build an accurate model of LLM capabilities if you wish to practically take advantage of their power while avoiding their pitfalls, which brings me to...

Last theme is the agent-native economy. The decomposition of products and services into sensors, actuators and logic (split up across all of 1.0/2.0/3.0 computing paradigms), how we can make information maximally legible to LLMs, some words on the quickly emerging agentic engineering and its skill set, related hiring practices, etc., possibly even hints/dreams of fully neural computing handling the vast majority of computation with some help from (classical) CPU coprocessors.

[Selvakumar R @R_Selva_kumar]

It's fun sipping coffee and let agents do the heavy lifting for you

[Selvakumar R @R_Selva_kumar]

@dwarkesh_sp This is the source for text becoming machine efficient. The road travelled

[Selvakumar R @R_Selva_kumar]

@dwarkesh_sp Text is machine efficient while videos, photos are costly. This is due to legacy architecture. The storage architecture would have been different if we didn't start with keyboard and started with camera as input to computer

[Selvakumar R @R_Selva_kumar]

@Jason I'm in