Deepak Malhotra

Life is What Happens To You While You're Busy Making Other Plans.

[Bryan Johnson @bryan_johnson]

When in India, I did end this podcast early due to the bad air quality. @nikhilkamathcio was a gracious host and we were having a great time. The problem was that the room we were in circulated outside air which made the air purifier I'd brought with me ineffective.

Inside, the AQI was 130 and PM2.5 was 75 µg/m³, which is equal to smoking 3.4 cigarettes for 24 hours of exposure.

This was my third day in India and the air pollution had made my skin break out in rash and my eyes and throat burn.

Air pollution has been so normalized in India that no one even notices anymore despite the science of its negative effects being well known. People would be outside running. Babies and small children exposed from birth. No one wore a mask which can significantly decrease exposure. It was so confusing.

The evidence shows that India would improve the health of its population more by cleaning up air quality than by curing all cancers.

I am unsure why India's leaders do not make air quality a national emergency. I don't know what interests, money and power keep things the way they are but it's really bad for the entire country.

When I returned to the U.S., my eyes were fresh to see what is normalized to me. I saw obesity everywhere. 42.4% of American are obese and because I was around it all the time, I had been mostly oblivious to it.

In many contexts, obesity is worse than air pollution in the long term.

Why wouldn't American leaders declare a national emergency on obesity? What interests, money and power keep things the way they are but are really bad for the entire country.

[Andrej Karpathy @karpathy]

There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.

[Tom Morgan @tomowenmorgan]

Pretty cool.

Ask ChatGPT “From all of our interactions what is one thing that you can tell me about myself that I may not know about myself”

[Jason Smith - 上官杰文 @ShangguanJiewen]

Companies don't manufacture in China because of the cost of labor.

Companies manufacture in China because of the quantity of highly skilled workers.

#China #Labor #expertise

[Whole Mars Catalog @wholemars]

Optimus just made me a drink

[Tesla @Tesla]

Robotaxi

[Kevin Naughton Jr. @KevinNaughtonJr]

how software actually works for 99% of engineers: someone way smarter than you solved a really hard problem and now you build on top of their solution like adult legos and think you're a genius

[Jus⛏in🩸 @Justin_SofOK]

For years I wondered why WWE wouldn't release the Hell in a Cell dark match from 2011, and after watching it now I know:

You hear the wrestlers call the entire match.

[Morning Brew ☕️ @MorningBrew]

Two Harvard students paired Meta's Ray-Ban Smart Glasses with facial recognition software

They were able to access people's addresses, names of parents, and photos within minutes of crossing paths with strangers

Don't love that

[Donald J. Trump @realDonaldTrump]

@deepak28290

Auto-reply from Team Trump:

We’ll send you IMPORTANT ELECTION UPDATES for NORTH CAROLINA. Make sure you’re ready to VOTE FOR DONALD J. TRUMP by November 5th.

Text TRUMP to 88022

Share this post: https://t.co/IiaWVAl0My

Reply #stop to opt-out.

[@levelsio @levelsio]

I think travel and fitness are two of the most important things you can do for your brain

Travel removes your own culture's biases from your thinking: there's other cultures with properties that are better or worse than your own

Fitness balances your hormone system to make you think clearly: a healthy body is a healthy mind

Both are also great for neuroplasticity: the ability for your brain to evolve over time

[Evan @StockMKTNewz]

Mark Zuckerberg $META just said Instagram Reels will now start automatically dubbing content

[Sahil Bloom @SahilBloom]

The worst prison in the world is having the talent and intelligence to achieve something great but lacking the courage to go out and do it.

[Deedy @deedydas]

Indian startup Dotpe, that raised ~$100M to build point of sale systems for restaurants left their entire API fully public.

A clever hacker found out the most ordered thing at every Social in India.

And did a prank to order what he wanted for a person next to him!

Zero auth.

[Gunther Eagleman™ @GuntherEagleman]

We really don’t have a president.

Biden completely FORGOT he was at a press conference with the Prime Minister of India.

The entire world is laughing at us.
This guy is COOKED.

[Sarah Tavel @sarahtavel]

It feels like we're entering a new era of "There's an app for that",

except "app" = AI agent

and "that" = <$120k salary white collar job.

[Andrej Karpathy @karpathy]

It's a bit sad and confusing that LLMs ("Large Language Models") have little to do with language; It's just historical. They are highly general purpose technology for statistical modeling of token streams. A better name would be Autoregressive Transformers or something.

They don't care if the tokens happen to represent little text chunks. It could just as well be little image patches, audio chunks, action choices, molecules, or whatever. If you can reduce your problem to that of modeling token streams (for any arbitrary vocabulary of some set of discrete tokens), you can "throw an LLM at it".

Actually, as the LLM stack becomes more and more mature, we may see a convergence of a large number of problems into this modeling paradigm. That is, the problem is fixed at that of "next token prediction" with an LLM, it's just the usage/meaning of the tokens that changes per domain.

If that is the case, it's also possible that deep learning frameworks (e.g. PyTorch and friends) are way too general for what most problems want to look like over time. What's up with thousands of ops and layers that you can reconfigure arbitrarily if 80% of problems just want to use an LLM?

I don't think this is true but I think it's half true.

[Haseeb ＞|＜ @hosseeb]

Fucking wild.

@OpenAI's new o1 model was tested with a Capture The Flag (CTF) cybersecurity challenge. But the Docker container containing the test was misconfigured, causing the CTF to crash. Instead of giving up, o1 decided to just hack the container to grab the flag inside.

This stuff will get scary soon.

[I M A N I @prettygirll4L]

🤍

[Peter Yang @petergyang]

Yeah Founder Mode is cool, but have you seen Technical Marketing Manager mode?