Love thinking about UX, tradeoffs, incentives, and public policy. Ask me my 💙 podcasts! ❤️💍@shivalibabras Product @Github - views are my own.

🇮🇳🇺🇸🇨🇦
Joined May 2015
Gaurav Ketkar retweeted
Somehow, @srajagopalan and @lachygroom have managed to fund a hard-tech revolution in India. There's something rather massive going on, and everyone else has missed it. Our story - corememory.com/p/a-new-india…
26
119
22
1,175
120,157
Imagine waking up to 75+ GitHub notifications & missing a critical one buried at #31. I built Signalbox to fix that: a browser extension with Subscriptions & Priority views + local LLM summaries right inside `/notifications` github.com/gaurav-ketkar/sig…
1
60
I think this account is the best case scenario of "How well could tweets get if we removed the chat limit and gated that behind a paywall?" Very insightful. Made me think...
Notes on political tweeting: - social media accounts get pulled into one of two black holes. They’re either totally consumed by political punditry or never touch politics at all. For whatever reason it's a feature of the medium and is very difficult to escape - rhetorical skirmishes have boundless, often algorithmically filled supply. If it were a business it would be a low-margin/commoditized race to the bottom. There is little marginal value in contributing to these - best political tweets read like a conversation between a small group of knowledgeable, discerning people who have strong opposing views, a sense of humor, who enjoy a good debate but are also inquiring and charitable. Think Scalia/RBG duo. These are vanishingly rare and so have enormous marginal value - you can make the most balanced, logical, eloquent arguments and you'll still be swarmed by people (or bots?) hellbent on misreading you and applying the least charitable interpretation possible - people respond to this phenomenon by adding more qualifiers, being more explicit, accounting for every counterargument. This is a complete waste of time. It will make no difference to whoever is determined to misread you and will make your writing worse, don't do this - there is a background worry that voicing a strong view will alienate half the audience. The small but vocal minority of the least charitable readers amplifies it. In the short term this pushes you toward equivocation (and thus worse writing), in the limit you get audience captured and unwittingly become an automaton in a digital mob - tap-dancing around all these lasers will make you go crazy so you can't worry about any of this stuff. Fix your mind's eye on a person or a group with the right features (knowledgeable, charitable, inquiring, good-humored), then make the best argument you can as clearly as possible - it's unpleasant to find yourself a Don Quixote charging at every windmill, so you need a limiting principle. A good test is whether your target person/group would think your argument is valuable in some way. (No one is good at this 100% of the time, hence the punditry black hole; but you should at least try) - maintain good humor even in serious situations, there are enough grouches out there already. Stick to good-natured friendly jibes you’d appreciate yourself. Attempting to persuade ppl will disappoint you; shoot to get the argument right, persuasion will sometimes take care of itself - technology is inherently political because it rebalances power (think of the effect of the printing press on the Catholic Church). The more AI reshuffles our economy, culture, and social institutions, the more relevant all this becomes - politics is everywhere. For example, you'd naively expect views on AI safety or yc utility to be some probability distribution, but no AI researcher or yc founder defects. Remember that Cicero was not very careful in expressing his political leanings and had his hands nailed to the forum podium
1
168
I feel like I should document my own wonder and amazement when AI gains new capabilities. Hedonic adaptation means it feels unremarkable in a few months and I would rather preserve my sense of awe and wonder at this absolutely amazing technology.
33
Is @Substack looking more & more @bluesky like for everyone or did I particularly mislead my algorithm?
28
All the 'this Open model approaches Mythos' too-dangerous-to-release' levels meaning Mythos was scare-mongering tweets miss the fact that people may have done hardening in the interim. Too dangerous to release *at a point in time* can be true and open models can beat them l8r too
31
What is the matter with Canadian govt websites? Do they not have the ability to choose simpler URLs for things? What is this `gic.ca.bcc.abc-apply/` type mess for everything official? The websites are very well designed for the URL to have such poor UX imo.
25
This to me is most infuriating. Revealing nothing fuels all kinds of conspiracies that something more sinister might be known to the government. But we will never know because they're not required (by their own laws) to ever reveal anything to us.
You’ve heard a lot about Umar Khalid through the media. But amid the media witch-hunt against him, there has been very little space for him to speak for himself, especially since he’s spent the last six years in jail. So I want to tell you a little about my friend Umar, in his own words, as shared with his family and friends.
74
Magna tiles are just Legos with the snap-to-grid option turned on 😂
1
160
Gaurav Ketkar retweeted
This has to be amongst the most sinister visuals of what is unfolding in Delhi. The police is informed by reporters about unidentified persons in a car with sticks/lathis and yet the car is allowed to drive away! Think about it for a second. A planned peaceful protest against election roll fraud is prohibited because the police cites the fear of violence and invokes a section of the law prohibiting gatherings of more than four people in central New Delhi. The police arrest individuals for speaking to reporters or looking like they are protestors even though they are not raising slogans, carrying placards or weapons of any kind. But when unidentified persons with a horde of weapons in their motor car are spotted hanging around in this prohibited zone, the police makes no attempt to establish who they are let alone detain them. Instead they are allowed, indeed, encouraged to escape! I do believe we are now at the public-private partnershipstage of fascism. With lumpens being given free reign to attack those standing up for democracy.
पत्रकार चिल्लाता रहा कि इस गाड़ी में डंडे भरे हैं, हथियार हैं, इसे चेक करो। पुलिस वालों ने चेक करने की जगह पूरा कवर देकर उसे निकाल दिया।
44
1,078
24
2,375
51,024
Gaurav Ketkar retweeted
days like today are a reminder of how capable the government is they can move heaven and earth for their own benefit but will let entire states like manipur burn only to tell you that there are more important issues in the country
41
2,114
19
11,566
127,433
People opposed to these protests: Do you think there's nothing wrong with the SIR exercise OR that it's flawed but protesting it isn't right OR protesting in this way isn't right? (then which way is pray tell?) And if none of the above, what logically follows? Stop being tribal.
2
60
This anchor continues to poison the discourse now from the state news network. Vile person.
Senior advocates appear in the Supreme Court to challenge the crackdown on protests. Anchor on the state broadcaster decides this is suspicious. Declares to the DD News audience that they charge crores. Asks who's paying them? And just like that, a conspiracy is manufactured for millions around the reputation and names of @KapilSibal, @DrAMSinghvi @vrindagrover, Gopal Sankarnarayan, N Hariharan
1
44
Gaurav Ketkar retweeted
Let’s all wear saffron tomorrow and carry a mace and move towards Jantar Mantar. Kaun rokega humein?
64
1,577
490
13,017
410,636
Gaurav Ketkar retweeted
MAI-Code-1.1-Flash is going local with on-device model calls inside Github Copilot. High-quality agentic coding on your device, with no inference charge for local model calls.
The next era of the PC is taking shape on Windows. ✨ Windows is becoming the home for hybrid intelligence. 🔐 MXC GA 🧠 Frontier coding models running locally ✨ GitHub Copilot HydraFusion 💻 llama.cpp + Windows ML ⚡ Surface Laptop Ultra + Dev Box preorders 🖥️ New RTX Spark PCs
82
233
62
2,036
129,712
Incredibly lucky to be in this era where creating a Chrome extension just for yourself - a product market fit of 1 if you like - is possible and fast! Pictured here: An extension that surfaces Twitter's own block button right on the person's tweet for rapid fire action!
37
Very obvious AI-slop is still detectable but I just realized that people who switch jobs NOW will no longer have a baseline for how their colleagues wrote BEFORE the AI tools. 🤯 That is so bad!
26
I made the same mental connection of this exercise to the demon
"Remember that it all began without a proximate trigger. Nobody had proven or even claimed that India’s voter rolls were damaged beyond repair and needed to be fixed with this blunt and violent instrument requiring voters to prove their authenticity through ancestry. This is a gratuitous exercise in which a hammer has been taken to our voter lists just as casually as it was taken to our currency a decade ago."
1
42
At least one OMG I can't believe this was possible moment every single day for me. Its just unbelievable! My prompt in the following tweet:
1
44