Vivek Singh (@petmongrels) profile picture

Understand - software engineering, product development, finance, public policy, and healthcare.

Bangalore
Joined March 2009
Photo by Vivek Singh (@petmongrels)Photo by Vivek Singh (@petmongrels)Photo by Vivek Singh (@petmongrels)Photo by Vivek Singh (@petmongrels)Photo by Vivek Singh (@petmongrels)Photo by Vivek Singh (@petmongrels)
One of the reasons most investors and journalists believe in AI bubble is because they aren't engineers, and hence cannot contemplate how impactful this technology is.
54
Professional and institutional investors are in business of growing AUM. Not XIRR of customers. Most important to know for anyone who gives money to them.
44
The UX challenge of the decade is to create something that combines terminal and browser in a single app.
1
1
71
1908: the Lancet calls for 18 age limit on reading in bed
24
119
33
462
318,131
This had to happen. I always believed Tukaram Munde standards are unimplementable uniformly. It is not chalta hai, it is majboori hai. Per capita income.
What India needs more of is "nahi chalta hai." From @TheEconomist
54
Without new technology there will be no GDP growth in countries that have fully distributed existing technologies. Lower GDP growth is still growth. And in absolute value growth it is quite material.
NYU Prof. Aswath Damodaran, the world's leading valuation expert and AI skeptic, on the ‘AI will triple GDP’ hype: “Real GDP growth has been lower globally over the last forty years than it was before PCs. All this stuff that claimed to make us more productive just redistributed wealth rather than make us all more wealthy.”
49
As a general rule, most "regrets" after one is caught's aren't regrets about having committed the action, but regrets about having been caught. Apologies count only when they are spontaneous, coming out of deep principles. Otherwise they are just an expression of submission.
57
790
51
3,396
My early impression of trying different methods, is similar. For engineers it is another set of tools to add and use where it fits best. Not a silver bullet.
while i'm quite bullish on formal methods, i think it's worth tempering expectations and doing your own research on it (disclaimer: i am not an expert in it). it cannot guarantee your code is bug free, only that it matches the spec under certain assumptions, and the specs come from what people want, not math. as long as we build software for humans, nothing can prove that your spec is right, only you can judge. i created the /statechart skill in pstack primarily because it helps make the state machines implicit in your code more obvious and legible to you, so you can shrink the review surface and review the work at a higher level: does it actually match your intent and goal? i used this skill a ton while building my game tibo.dog, below you can see the statechart for story mode where i was able to find many bugs with it. in that sense, i think model checking (eg TLA+ with TLC) is quite useful for understanding and checking properties of your system e2e, and you don't have to write tons of new verification code for it. the spec isn't your actual code so there's always a risk of drift, but so far i've had good experiences finding actual bugs with it. i tend to pair it with empirical verification (running the code) also to confirm it's real other types of formal verification like proving theorems (eg Lean, Dafny), bounded model checking (eg CBMC, Kani) etc have their own tradeoffs. proofs cover every possible input, but only for the properties specified, and it takes a lot more code/work. bounded model checking can run on your actual code, which avoids drift, but it only checks up to a certain number of steps, so it's better at finding bugs than proving there aren't any so i think in the near-medium term you'll still need to layer your codebase: constraints and guardrails (types, lint rules, api design, refactoring), empirical verification (tests, evals, real usage), and formal verification for the parts that really need it. none of these catch everything on their own, but layered together, they catch a lot more and shrink what's left for you to review but also maybe in the limit, the bitter lesson will prevail and flawless code will just be printed out perfectly every time with no bugs. even in that scenario, /statechart still seems very helpful in confirming that the agent understood your intent correctly. give it a try if you haven't already! github.com/cursor/plugins/tr…
1
75
The main drawback of technical analysis is that is not very useful.
This quoted post is unavailable.
72
As a customer of a financial institution I cannot legally do penetration testing, for my own due diligence and peace of mind. 😃
A Stockal Investor Says Her Long-Term Investments Were Sold Without Consent. Here's Her Story Speaking exclusively to NDTV Profit, the investor recounts how alleged unauthorised trades unfolded, contributing to claims of more than ₹100 crore in losses across affected accounts. @TamannaInamdar @niraj_shah @alexandermats
74
There is blessing in this choice. Jobs.
If you use inferior models, home grown harness, and app - with each choice you lose a significant % productivity. You may be at 50-100% disadvantage just based on these choices. It may get levelled in future, but that is few years away.
70
If you use inferior models, home grown harness, and app - with each choice you lose a significant % productivity. You may be at 50-100% disadvantage just based on these choices. It may get levelled in future, but that is few years away.
By building all three, i.e. model, harness, and app - Claude, especially Claude Code, is setting itself to be a premium offering like Apple did with iPhone. "Don't think too much just use Claude"
1
113
Vivek Singh retweeted
classic lant pritchett, vindicated yet again.
6
165
15
970
54,694
Foldable phones may do much better than iPads / tablets
33
dot in domains require KYC, so for temporary use cases if I want a domain - I see that .in is the cheapest domain. Even cheaper that .xyz But I don't want it. Hence probably it is cheap.
44
By building all three, i.e. model, harness, and app - Claude, especially Claude Code, is setting itself to be a premium offering like Apple did with iPhone. "Don't think too much just use Claude"
1
1
97
The UX keeps growing on you, sometimes because of model, or app + harness.
28
If you listen carefully you can hear the AI bears lifting the goalposts and moving them another 100 yards down the field after Anthropic's S1 was released.
32
18
5
690
44,897
How do you value something that gives three growth data points: 12x, 14x, 5x What do you as next growth rate: 4x, 2x, 1.5x, 0.75x, 6x? Anyone putting even very 75% growth thinks it's bubble. Very difficult to value very high growth stocks.
1
47
Vivek Singh retweeted
We got live dosa tracking in Bangalore before GTA 6
66
67
32
1,761
75,312