My strategy for getting a job is: 1. contribute to the company's open source project (often aggressively 😅) 2. make an impact with my contributions and 3. ask for a job to do this full-time So far, it has worked EVERY time.

The combined essays of a million non-contributing bystanders is worth less than a ketchup stain on a mcdonalds napkin from someone who is an active participant in a project. Open source is shaped by contributions and participation, and failing that by forks and competition.

... chances of winning, but people on Polymarket never chose to bet against India. It's also easy to see that India was the favorite going into the tournament. Or maybe there could be a bias in the holders.

I'm curious to see if other predictions over time during the match day had similar trends or whether it's just that you never bet against India's 11 for the tournament. Like statistically, India can lose the match, and there might have been points where Pakistan had more ...

Asia Cup, India vs Pakistan on Polymarket. Notice the fluctuations during the match, but how it never dropped below 60:40.

I'm coming prepared to @IndiaFOSS this year. This is the most I've been excited for a conference in a long time. I'm particularly looking forward to the public policy-FOSS overlap, which is very unique to @FOSSUnited, and I feel has been growing in the last ~3-4 years.

For a moment, I was like, "When did we have an Indian GP?"

You know, I'm something of an analyst myself. Pazhamporis on me, guys!

Stupid statements like this may sound good because people often judge policies based on their intentions rather than the actual consequences. As they say, "the road to hell is paved with good intentions."

OpenRouter down. Time to remember what sunlight feels like.

Private AI will be a massive push over the coming years Freysa team getting ahead of it and shipping Higher

We've been working on improving privacy while using LLMs. Think about how much ChatGPT or any of these LLMs (and the companies behind them) already know about you. We’re building systems where you get the best models without having to hand over your identity.

1/ Launching Enchanted ✨ Every human deserves a truly personal AI. Not one controlled by a big lab. Not one that hoards your data. Use Enchanted. link.freysa.ai/enchanted

Finally open source maintainers have a way to pay contributors with all the money they are making.

I fried my MacBook's "logic board" and the service center guys say I have to replace the entire thing? Including the CPU and the SSD? WTF? And it costs as much as the MacBook itself! I'm pretty sure the whole thing isn't gone, and some unaffiliated technician can fix it at the fraction of the price. This might be the tipping point where I go full reclusive Arch Linux hermit.

Despite what the internet says, it's safe to assume people from anywhere are friendly.

This insight explains OpenAI's acquisition of Windsurf. From "Laws of Tech: Commoditize Your Complement" by Gwern

the flood of spam messages on slack by people who just join slack channel 1-2 days prior and question how the xyz people are selected in open-source programs like GSoC is really sad and i am pointing this out here coz ppl especially who are new to open-source or just join open-source for tags like GSoC and all and when they arent selected they spam the slack channels asking the credibility of ppl who are selected and say favouritism and what not, i would say get your facts right, see the ppl’s contributions and GSoC is not a program for beginners its a program designed to reward existing contributions and foster open-source collaboration. The indian tech community is definitely not defined by such dumb ppl who argue pointlessly but its definitely defined by experienced folks who have done tremendous job in the community. So if you are new to open-source, please understand that it requires patience and consistent efforts to get the good things and success, coz success always comes after u put consistent efforts! Peace✌️

I have been thinking about this every day since I first read it. From "Make Something Heavy" by @anuatluru

Some career advice: there are only 3 things you can optimize for when taking a job. 1) Money. 2) Learning. 3) Fun. In my experience it’s very rare that you have all 3: you should optimize for 2, at best. Depending on where you’re at with life and career, you might heavily optimize for one over another. Sometimes, you’re making a crap ton of money and learning a lot but the hours suck, your on call constantly, and you’re just not having a ton of fun. Sometimes, you’re riding the fun wave with your awesome coworkers playing ping pong every day while growing intellectually but you’re making very little dough. Sometimes, you’re making lots of money and having tons of fun doing it but working on the most boring, monotonous things that you’re not learning anything useful from. A failure mode I’ve seen is optimizing for ONLY 1 of these. If you’re having tons of fun but make nothing and learn nothing, you’re basically Bighead and at risk of complete career stagnation. If you make shit loads of money but hate your life and are learning nothing at all, you’ll end up absolutely miserable regardless of how flush with cash you are. And endless learning with no money or fun is no different then slogging through academia, regretting your life decisions. Be warned: there are LOTS of other things employers try to sell you on but are mostly a gamble and not worth optimizing for: 1). Promises of power This usually comes across as “impact” or “leadership opportunity” or “influence”. And, yes, while you can learn a lot from a good leadership opportunity, power is not something anyone truly wields and is very rarely bestowed upon that employer to give out: it’s something handed down from greater forces (management, industry trends, VC group chats, the markets, executives, etc.) and is easily stripped away. Optimizing for promises of power is a dangerous game climbing the endless corporate treadmill. If you really want to roll the dice on “impact”, it’s probably better to go start your own company. 2). Titles At a certain point, all titles are completely meaningless. There may be a meaningful distinction between a junior engineer and mid level engineer, but there’s basically no meaningful difference between a senior staff engineer and a staff engineer besides who management likes better, who got luckier in the promo cycle, luck, or timing. Optimizing for a title is a very bad idea beyond what it might do for your career in the short term: if you’re having no fun, making no money, and learning nothing, it doesn’t really matter if you made “staff engineer” (but it might help you go land a better job somewhere else). 3). Bubbles Don’t go into something that’s REALLY hot in the moment (like AI, crypto, web during the dot com burst, etc.) just because it’s trendy: you’re gambling on the future of that technology and, historically, everyone everywhere (including you) are bad at predicting the future. 4). Technologies Unless you’re heavily optimizing for fun, don’t pick a company just because of the tech they use. There’s a reason Java devs are still in demand: not super fun or stimulating technology but every corporate entity uses it to some degree and those skills will probably always be in demand (so will likely make you money and is worth staying sharp on - but again, probably not very fun). There’s a graveyard of technologies (that people had to spend precious time learning) and products that made no money and probably weren’t worth learning.

This is not different from regulatory capture or the thing we blamed US companies for when they wanted to regulate AI. India has a unique opportunity to do something different before the big tech players realize the value in having non-English AI models but we continue to resort to Babudom. Do I need an address proof to access your AI?