Sri Prasanth S

Serial entrepreneur. Gaming CTO, FMCG CTO & COO.

[Andrej Karpathy @karpathy]

Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.

- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.

The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.

So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.

[Elon Musk @elonmusk]

Worth reading about this

[Pritam Chauhan @Im_pritam18]

Dr. Pratosh at IISc Bengaluru tells his students that rapid advances in AI are commoditizing intellectual labour.

He raises an unsettling question: if companies stop recruiting on campus in the next 5–10 years—even at India’s top universities—what will the purpose of a university education be ?

A must watch video for everyone in tech

[Andrej Karpathy @karpathy]

One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.

[Science Simplified @Scivf4]

The person who invented the sewing machine is a genius…🤯

[Sri Prasanth S @sriprasanthme]

Claude is down! Time for me to take a break!

[Ihtesham Ali @ihteshamali]

Broadcom's CEO just told investors that one senior engineer with Claude Opus can do in one week what ten engineers earning $300,000 a year take three months to produce. Every engineering org in the world is running the wrong headcount math.

This was not a prediction. It was a report from inside Broadcom.

Hock Tan said his engineers are already using Claude Opus for engineering design and code assistance. Not experimenting with it. Using it. And the output they are getting from a single senior engineer in one week is equivalent to what a team of ten would produce over three months.

Do that math slowly.

Ten engineers at $300,000 a year is $3 million in annual salary. Three months of their time is $750,000 of labor. One engineer with Claude Opus produces the same result in one week.

He also said something that most people in the room let slide past them.

The engineers who get good at these tools do not just get faster. They get better. The designs they produce with AI assistance are probably better than what they would produce without it.

That is the part nobody wants to say out loud.

It was never just about speed. The ceiling on what a single engineer can produce just moved. And the teams that have not figured that out yet are spending $750,000 to do what one person with the right tools can do for a fraction of that.

What do you guys think about this?

(Watch the full interview at @BloombergLive on YouTube)

[Dreams N Science @dreamsNscience]

How Jet Engines Create Massive Thrust 😳

🎥 sscientific.girl

[Sri Prasanth S @sriprasanthme]

Elon Musk’s current net worth is sitting at around ₹123 lakh crore. To put that in perspective, if he decided to split his money equally with every single person on Earth, we'd all get roughly ₹15,350. Someone pass this on to Elon, my UPI ID is ready. 😭💀 #ElonMusk

[Sri Prasanth S @sriprasanthme]

Get your Idiot Index calculated and understand your efficiency in your manufacturing process.

[Eric Jorgenson 📚 ☀️ @EricJorgenson]

Everyone can use @elonmusk's "Magic Wand Number" and "Idiot Index"

They're universal ideas, helpful in any industry.

[Thariq @trq212]

been asking others at Anthropic how they stay in the loop with Claude and fully understand the work being done

this is one of my favorites from Suzanne:

[Marques Brownlee @MKBHD]

Yeah Wemby is good and everything, but he’s not the only ridiculously skilled person in this arena 🤯

[Science Magazine @ScienceExpand]

Physics within geography

[Alexander Whedon @alex_whedon]

Introducing SubQ - a major breakthrough in LLM intelligence.

It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),

And the first frontier model with a 12 million token context window which is:

- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus

Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).

Only a small fraction actually matter.

@subquadratic finds and focuses only on the ones that do.

That's nearly 1,000x less compute and a new way for LLMs to scale.

[Sarvam @SarvamAI]

We are excited to announce that Sarvam is partnering with @PixxelSpace to power the AI backbone of India's first orbital data centre satellite.

This is a first for the country, with India-built AI models running on an India-built satellite and both training and inference happening directly in orbit, without any dependence on foreign cloud or ground infrastructure.

[Sri Prasanth S @sriprasanthme]

@Yahoo you really need to work on the monthly and yearly pricing for India if you're serious about selling storage space in Yahoo Mail. Looks like basic testing is missed before the release.

[Sri Prasanth S @sriprasanthme]

DeepSeek v4 just answered the car wash question sensibly. The first open-source model to do so. https://t.co/kfttMLnYfP

[NetworkChuck @NetworkChuck]

Gemma 4 running on my iPhone works without internet, is blazing fast and can translate Japanese from a pill bottle. 

Local AI models running on a phone feels like magic.

[Andrej Karpathy @karpathy]

LLM Knowledge Bases

Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So:

Data ingest:
I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them.

IDE:
I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides).

Q&A:
Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale.

Output:
Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base.

Linting:
I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into.

Extra tools:
I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries.

Further explorations:
As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows.

TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.

[Sri Prasanth S @sriprasanthme]

The barrier to building enterprise software has collapsed.

One person. AI tools. Domain knowledge. That's all it takes.

Built an entire internal product suite for a manufacturing business — solo — for under $10/month.

→ https://t.co/Oex9ndrKIJ