Sharad Regoti

@signulll People at OpenAI/Claude think they are building God, whereas at Google it's just a job for them. That is the difference.

google right now has almost zero cultural gravity when it comes to frontier ai discussions in any of the conversations i am in, like not even mentioned once. both in consumer + enterprise contexts (i naturally tell ppl to claude it or chatgpt it instead of googling it, i can’t recall the last time i told someone to google it). for our startup, we used to use gemini for some things cuz the price of some of the flash models but even that now makes very very little sense (we started using openai & it’s been great). in essence almost all of the frontier discourse is owned by openai/anthropic, & all of the open weight momentum has shifted entirely to aggressive chinese labs. i don’t know where google fits in. i guess when you don’t have an existential threat anymore, there is no reason to be aggressive.

@IndianTechGuide Exchange is defined as "A two-way give and take between two people or groups" The correct word here would be "Transfer" or "Replace"?

🚨 India and Japan announce a plan to exchange over 5 lakh workers in five years, including 50,000 skilled Indians.

@warikoo I like this take

My prediction of how AI will change the way we work 1/ Most knowledge work will be done by AI agents, not humans. The spreadsheet, the report, the code, the first draft of everything. 2/ Companies will pay for these agents like they pay salaries today. Monthly, recurring, per seat of work done. 3/ What will make the agent valuable and paid more is what its human owner feeds it: their data, their judgment, their relationships, their reputation. A CA's agent is valuable because of 20 years of the CA's tax judgment inside it, not because of the AI. 4/ Companies will pay for agents the way they pay for employees today, but the pay will follow trust, not just capability. Raw capability will be cheap and everywhere. Accountability will be scarce. The human who says "I stand behind this agent's work" will be the actual product. 5/ Income will follow a creator-economy curve, not a salary curve. The power law as we call it today. A small number will build agent empires. A larger middle class will earn well in narrow niches, the way a good local YouTuber earns today. Most will earn a little. 6/ Because most will earn only a little, governments will be forced to tax the agent economy and redistribute it. Let's call that Universal Basic Income. There is Elon's version as well (quite plausible) that the speed of good/services will far outstrip speed of money supply, so money can be printed, given to everyone (in abundance) and it still wouldn't cause inflation. 7/ People will still work, but the work changes shape. You stop being the worker and become the trainer, the editor, the quality controller of your agents. You work on the machine, not in it. 8/ The winners won't be the best builders. Building will be easy. The winners will be the ones with something real to put inside the agent: rare knowledge, rare data, rare trust, or distribution. AI multiplies what you already have. Multiply zero and you get zero. 9/ "With real humans" tag will become premium. Be it travel, be it food, be it work, be it relationships. So the practical advice for an ordinary person is not "learn to build agents." It is "become excellent at something narrow and real" The tailoring, the surgery, the tax filing, the teaching, the sales, the tennis. And then bottle this excellance. The agent is the bottle. You are still the wine :))

@priteshlakhani Koreans were given...

Can someone dumb down exactly what is happening in Korea for someone who’s not a stock market expert.

@Polymarket Why can't you use simple numbers like 2 billion?

BREAKING: South Korea’s stock market crashes 44% in just 40 days, wiping out ~$2,000,000,000,000.00 in market value.

@buger at any given time, do you only work on a single task? I tried working on multiple tasks with agents, but I couldn't accomplish any of them; context switching kills everything. Now I only do one task at a time and read a book while the agent works.

People keep asking what the heck I'm doing with 30 billion tokens a month. Simple: an inhumane amount of work. Think about it — the whole point of the new agentic era is leverage. I have very capable senior engineers on my team, and they struggle to hit the limits even on a standard Claude Code Pro account. Not even talking about the $200 one. Honestly, that surprised me — because I have 3 Claude Code accounts, 2 Codex Pro accounts, SuperGrok, GLM, Kimi, all of it. And I consume every one of them until it hits the limit. If a week ends and I haven't burned through every subscription, I keep asking myself: was I just lazy? How can I push it more? What else could I automate? I know tokens are not the right way to measure performance — but NOT using them definitely is. And when I do hit all the limits, I just turn my brain off and enjoy life. Why aren't these engineers doing the same? Because they're trying to do what they did before, just with AI now. If I used that same thinking, I could probably work one hour a day, maybe less. Sounds plausible — but for me, it's not the answer. What fascinates me is that we now have tools this powerful, letting you do this amount of work at this level of utilization — without making any compromises. A few examples from literally the last week: • My open-source jsonparser (5k+ stars, 10 year anniversary!): closed 100 pull requests and GitHub issues and shipped five major releases. All while drinking coffee, essentially, in a matter of a few days. • rsync: I'm one of the people driving the next rsync release. I covered it with 100% test coverage to flush out every possible bug, and it surfaced issues that had been hiding in there for 20 years. • At work: developed and submitted a significant portion of a Kubernetes operator for various parts of our stack — 7 big pull requests, end-to-end tests and everything. • Proof: kept pushing my own product forward, on top of everything above. How? I run five agents in parallel. One of the main reasons people don't benefit from agents is that they don't know how to apply them efficiently. Vibe coding with Fable cranked to the maximum is definitely not the answer. Sometimes it is — but it's a small, teeny part. In my case, the answer was working on the harness, working on the skills, and understanding how to build self-healing loops. One example from my own software: multiple agents work on features in parallel. When they find an issue, they file it in GitHub automatically. Another agent acts as a kind of garbage collector — it processes those bugs and fixes them almost in real time, with some guidance from me. And my agents don't stop when I do — they work 24 hours a day. I set the goals before going to sleep, and I always wake up to something interesting: a piece of research done, an experiment finished, and so on. It's freaking amazing. And if I had 10x more tokens, I'd spend them all just as efficiently. So many ideas, so many experiments I need to do!

@digitlartifacts We have come a long way

women hand weaving memory for the early computers used on apollo lunar missions

@mainbhiengineer Europeans have earned it, we have not

Infosys was fined in France over tracking employees working time. France has only 35 hours of working week and any overtime needs to be paid and recorded. Compare it with our labour laws, Narayana Murthy comes on TV and say that people should work 70hrs a week and get away with that. Forget about Narayana Murthy, when our PM Modi is celebrated as working for 18hrs a day and no one questions it then you know what kind of toxic work culture and pressure we are promoting.

@mxtaverse Not true, I asked Groked "5 year jail for adultery only for men" This is gender neutral

>"uniform" civil code >tribals exempted >5 year jail for adultery only for men 😂🙏

@Polymarket Isn't this a security Nightmare? I hope they have thoroughly scanned their code using Mythos, etc...

JUST IN: Elon Musk announces 𝕏 will open-source its entire codebase.

Anthropic after this announcement

We got a working PS5 emulator before GTA 6

@rajarshidattapy It's all layers, built on top of each other

Gm.

@samyok Trust me benchmark

guess the benchmark

@jonathan_wilke Yes, you are the only one

Am I the only one who thinks git worktrees are fucking confusing?

@prakdadlani Do you have a personal blog or any content on how you started manufacturing in India, how you picked up a problem, how you financed your operation, or problems you have solved on the way? bureaucratic processes, corruption,

@SharadRegoti Silvassa

The easiest way to build a brand in India is to import. The hardest way is to manufacture. That's exactly why we're doing it.

@chandrarsrikant @aravind as you predicted

🚨 Exclusive: Deloitte, EY, KPMG among 7 firms in race as Centre prepares nationwide audit of critical IT infrastructure The Centre is set to announce a nationwide cybersecurity audit of critical government IT infrastructure in the coming days, with the process of selecting the agencies that will carry out the exercise entering its final stages. A senior government official told Moneycontrol that the initiative is "at the final stages" and an announcement is expected "in a couple of days." Documents reviewed by Moneycontrol show that NIC Services Inc. (NICSI), the government's technology procurement arm, has completed the technical evaluation of bids, shortlisted seven firms for the next stage of the process, and currently financial evaluation is underway. Records show that a government committee found that Qseap Infotech Pvt. Ltd., Ernst and Young LLP, AAA Technologies Ltd., AKS Information Technology Services Pvt. Ltd., KPMG, Deloitte Touche Tohmatsu India LLP and SecureyeS Techno Services Pvt. Ltd. had qualified at the "Eligibility & Technical compliance stage By @AihikS moneycontrol.com/technology/del…

@Ayush_cg As Sam Altman said, it will become a utility. Just wait and watch. Currently, we are limited by compute capacity and energy; once those are stabilized, this will become affordable.

Experts 2 years ago predicted models would get cheaper and cheaper. The opposite has happened. Frontier models keep getting more expensive per token, and agents are increasingly incentivized to run things in parallel. I'm pretty sure even these prices are subsidized.