Lightnews — Scholar-powered news

LightNews

@paraschopra.com

55 followers 690 following 690 posts

Posts Media Videos Starter Packs

paraschopra.com @paraschopra.com · 9h

What will be scarce in the AGI era?

that's what will make real returns 💸

paraschopra.com @paraschopra.com · 11h

In the age of LLMs, smart people are getting smarter, while dumb people are getting dumber.

paraschopra.com @paraschopra.com · 1d

via docs.google.com/presentatio...

Advice for a young investigator in the first and last days of the Anthropocene

Advice for a (young) investigator in the first and last days of the Anthropocene Jascha Sohl-Dickstein Anthropic Title: Advice for a young investigator in the first and last days of the Anthropocene Abstract: Within just a few years, it is likely that we will create AI systems that outperform the...

docs.google.com

paraschopra.com @paraschopra.com · 1d

Do you agree with this?

paraschopra.com @paraschopra.com · 1d

AI slop is like greenhouse gas emissions.

For personal gain, everyone is polluting the Internet - classic tragedy of commons!

paraschopra.com @paraschopra.com · 1d

You know this is true even for major life events.

Amputees readjust to their previous levels of happiness, and so do lottery winners.

this doesn’t mean happiness can’t be impacted; this just means the lever to it might not lie in achieving or losing things.

paraschopra.com @paraschopra.com · 1d

I think hedonic adaption is one of the deepest truths of life.

Its direct implication is that if you aren’t happy today, you won’t get happier in future when you acquire that shiny new trophy.

Seek reasons to be happy no matter where you’re in life!

paraschopra.com @paraschopra.com · 2d

Good mental health is a lifelong journey.

Because we observe people’s behaviours, never their thoughts, it can appear that everyone else other than you has figured it out.

But the truth is that everyone is battling their inner demons almost all the time.

paraschopra.com @paraschopra.com · 2d

What’s your personal benchmark for declaring that AGI has happened?

paraschopra.com @paraschopra.com · 3d

It seems like most successful entities (people, companies) have one trick up their sleeves that they keep milking for all its worth

Apple with their iPhones, Darwin with his theory of natural selection, Picasso with his cubism

This shows how rare breakthrough success really is!

paraschopra.com @paraschopra.com · 3d

You should make notes on everything you learn.

First, the act of making notes forces your brain to understand the topic much better than passive reading.

Second, written notes allow you to stitch a web of ideas and discover new connections b/w previous thoughts and new ones.

paraschopra.com @paraschopra.com · 5d

Non-fiction books should have a pyramidical structure.

One page summary at the start. Then few pages elaboration of it. Then progressive deeper dive into topics.

Authors should enable readers to focus precisely on areas readers want to while ignoring the rest.

paraschopra.com @paraschopra.com · 6d

From this lens, almost everything collapses to having near zero impact on how my life will turn out to be.

It’s crazy how we often make things seem way worse than they are by simply brooding over them.

paraschopra.com @paraschopra.com · 6d

The fastest way to turn a minor inconvenience into a major suffering is to think about it repeatedly.

Thumb rule: allocate thinking cycles only in proportion to a situation’s future impact on the quality of my life.

paraschopra.com @paraschopra.com · 6d

read it here: www.wittgensteinproject.org/w/index.php...

paraschopra.com @paraschopra.com · 6d

What a sublime essay!

the only “popular” lecture that Wittgenstein ever gave, it expressed how the mystical feeling about existence of universe is inexpressible in language and that’s what makes it “absolutely” valuable.

paraschopra.com @paraschopra.com · 6d

Whenever a yearning to answer the meaning of life arises, reinterpret it as your brain seeking the next interesting project it can engage with.

This yearning is a function of our dopamine system wanting more than it has and seen so far in life.

paraschopra.com @paraschopra.com · 7d

Plus, the level of control on videos is minimal and good creators want that.

I do think OpenAI will try acquiring TikTok in the US, and this is a step towards that.

paraschopra.com @paraschopra.com · 7d

Tried Sora on a friend’s phone, and it was… bizarre!

My prediction: in its current form, it’ll fizzle out

People want to see other people, & that’s why social media works. This infinite feed of AI slop feels like an artificial sweetener, nice at first but ultimately disgusting

paraschopra.com @paraschopra.com · 7d

Yeah, I know the irony of posting this on an infinite feed but where else would you post warnings about things that are harmful?

paraschopra.com @paraschopra.com · 7d

Read things that engage you deeply, my friends.

Once you start viewing infinite feeds as drugs, it becomes possible to start avoiding them altogether.

I now can’t stand an endless feed for more than a few minutes. My brain refuses to process that rot.

paraschopra.com @paraschopra.com · 7d

The wisest people are those who stop competing with others, and start seeing life as an unserious play where it doesn’t matter who wins and who loses.

(I think everyone arrives at this point eventually; some sooner, some closer to death)

paraschopra.com @paraschopra.com · 9d

- Then ask it to give feedback on my notes
- Enrich my notes

paraschopra.com @paraschopra.com · 9d

How I read dense technical papers:

- Read it on my ReMarkable (make lots of notes)
- Summarize it on Zotero (key ideas + what's not clear to me)
- Upload PDF on chatgpt and ask it to explain + make a strong case for it + criticise it
- Discuss what's unclear

paraschopra.com @paraschopra.com · 9d

People think RL is magic, but RL only works when you have near-infinite compute and a lifetime to collect diverse samples from the environment.

Humans obviously learn under finite compute and finite sample regime, so it’s worthwhile to learn how we do it.