Paras Chopra's banner

Paras Chopra

@paraschopra • 240,072 subscribers

life is a game 🕹️ • building @lossfunk

Shorts

Starlink satellites in space. Holy cow - these are too many! Didn’t realise there is a swarm up there.

Starlink satellites in space. Holy cow - these are too many! Didn’t realise there is a swarm up there.

880,638 просмотров

This is GPT-6 Astra playing VizDoom at 35 fps. How? Isn't the model slow to do any realtime control? Instead of playing the game directly, I asked it to train a small CNN (12k parameters) + code to play the game, observe failures and iterate until it could win this level.

This is GPT-6 Astra playing VizDoom at 35 fps. How? Isn't the model slow to do any realtime control? Instead of playing the game directly, I asked it to train a small CNN (12k parameters) + code to play the game, observe failures and iterate until it could win this level.

16,169 просмотров

Installed a bird camera in my balcony and captured this red bulbul in a frame.

Installed a bird camera in my balcony and captured this red bulbul in a frame.

79,809 просмотров

I gave Astra a comic book cover to animate in threeJS and this is what it came up with. So insanely cool!

I gave Astra a comic book cover to animate in threeJS and this is what it came up with. So insanely cool!

11,679 просмотров

I have been using Wispr Flow, and it seems like a game changer. Just press Fn and speak - that's it.

I have been using Wispr Flow, and it seems like a game changer. Just press Fn and speak - that's it.

47,620 просмотров

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

I may have found two new Earth-sized planets in our galaxy this morning! Inspired by yesterday's reddit thread on how someone used Claude Code to find a new planet, I spent this morning searching NASA TESS data on nearby stars that TESS scanned less thoroughly (at low cadence). To my surprise the pipeline discovered two previously undocumented exoplanet candidates: • TIC 370612437 (Cygnus constellation, 158 light-years): dims by 0.02% every 3.56 days → ~1.0× Earth's size. • HIP 114417 (Andromeda constellation, 134 ly): 0.03% every 4.93 days → ~1.1× Earth. The method to find them is to search for periodic dims in light coming from a star when a planet passes in front of it. For both the candidate exoplanets, the dip repeats 55 and 21 times respectively across 4 separate observing years (2019-2024). These are obviously tentative and unconfirmed, but several signals point to them being credible candidates: • same dip in NASA's and MIT's separate processing of the images • hide one year's dip and predict its times from rest of the dips (matched in 3 of 4 years) • flip light curves to create fake signals and validate observed dips against it The method and results were reviewed by independent agents: Codex and fresh Claude. Important caveat: a faint neighbouring pair of eclipsing stars could fake the dip, so these exoplanets remain candidates pending further verification. What's next: report to NASA's ExoFOP as Community TESS Objects of Interest, so astronomers can verify. The next scan of these stars via TESS will happen in Aug 2027 - Apr 2028, if the preregistered dip predictions hold up, these are indeed strong candidates. Nicknames for these exoplanets: Aakanksha (Sanskrit: aspiration) and Orbit (well, Orbit)

I may have found two new Earth-sized planets in our galaxy this morning! Inspired by yesterday's reddit thread on how someone used Claude Code to find a new planet, I spent this morning searching NASA TESS data on nearby stars that TESS scanned less thoroughly (at low cadence). To my surprise the pipeline discovered two previously undocumented exoplanet candidates: • TIC 370612437 (Cygnus constellation, 158 light-years): dims by 0.02% every 3.56 days → ~1.0× Earth's size. • HIP 114417 (Andromeda constellation, 134 ly): 0.03% every 4.93 days → ~1.1× Earth. The method to find them is to search for periodic dims in light coming from a star when a planet passes in front of it. For both the candidate exoplanets, the dip repeats 55 and 21 times respectively across 4 separate observing years (2019-2024). These are obviously tentative and unconfirmed, but several signals point to them being credible candidates: • same dip in NASA's and MIT's separate processing of the images • hide one year's dip and predict its times from rest of the dips (matched in 3 of 4 years) • flip light curves to create fake signals and validate observed dips against it The method and results were reviewed by independent agents: Codex and fresh Claude. Important caveat: a faint neighbouring pair of eclipsing stars could fake the dip, so these exoplanets remain candidates pending further verification. What's next: report to NASA's ExoFOP as Community TESS Objects of Interest, so astronomers can verify. The next scan of these stars via TESS will happen in Aug 2027 - Apr 2028, if the preregistered dip predictions hold up, these are indeed strong candidates. Nicknames for these exoplanets: Aakanksha (Sanskrit: aspiration) and Orbit (well, Orbit)

55,275 просмотров • 9 дней назад

Claude 5.5 is crazy good with animations. Following is one-shot journey from a single bacterium cell all the way to quantum fields constituting it. Awe-inspiring! (It chose to add an original score to it too!)

Claude 5.5 is crazy good with animations. Following is one-shot journey from a single bacterium cell all the way to quantum fields constituting it. Awe-inspiring! (It chose to add an original score to it too!)

73,016 просмотров • 24 дней назад

My favorite way of interacting with Claude Code is to have it generate static HTML files as outputs (reports, explorations, code structure, mockups etc.) I wanted to iterate on the file by commenting in browser and having Claude update the output live. So, I built this Claude Skill👇 How it works: - Install Claude Code skill (ask it to clone repo) - Build an HTML page for anything (e.g. research coding agents and generate HTML report) - Ask it to make the page interactive That's it. CC will launch a localhost server and allow you to then leave comments on the page itself and once it updates, will give you a tour of changes. It's like Google Docs kind of comments/iteration but for HTML pages.

My favorite way of interacting with Claude Code is to have it generate static HTML files as outputs (reports, explorations, code structure, mockups etc.) I wanted to iterate on the file by commenting in browser and having Claude update the output live. So, I built this Claude Skill👇 How it works: - Install Claude Code skill (ask it to clone repo) - Build an HTML page for anything (e.g. research coding agents and generate HTML report) - Ask it to make the page interactive That's it. CC will launch a localhost server and allow you to then leave comments on the page itself and once it updates, will give you a tour of changes. It's like Google Docs kind of comments/iteration but for HTML pages.

216,699 просмотров • 4 месяцев назад

Created a virtual bookshelf of all books that I've read and reviewed online! A total of 388 books, from 2014 to now. Link below 👇

Created a virtual bookshelf of all books that I've read and reviewed online! A total of 388 books, from 2014 to now. Link below 👇

30,671 просмотров • 1 месяц назад

I tried replicating fully agentic coding on my laptop and I'm impressed! My setup - Qwen 27bn (4bit quant) - Pi coding agent - CCO sandboxing for yolo mode My Mac M3 has 36GB ram. It's surreal to see a local model read prompt, follow it flawlessly and then build a fully self-contained html/css page with zero errors. I got about ~20 tokens/second, and on this prompt quality similar to Haiku! But... this is running locally :)

I tried replicating fully agentic coding on my laptop and I'm impressed! My setup - Qwen 27bn (4bit quant) - Pi coding agent - CCO sandboxing for yolo mode My Mac M3 has 36GB ram. It's surreal to see a local model read prompt, follow it flawlessly and then build a fully self-contained html/css page with zero errors. I got about ~20 tokens/second, and on this prompt quality similar to Haiku! But... this is running locally :)

107,538 просмотров • 5 месяцев назад

How LLMs actually work! Recently gave a 2-hour long talk on this. 42 slides, all yours 👇

How LLMs actually work! Recently gave a 2-hour long talk on this. 42 slides, all yours 👇

75,382 просмотров • 7 месяцев назад

Doing #30daysofpushup challenge. Day 1/30. Want to join me? Post your video. If you publicly commit, you’re more likely to see through the 30 days and then it’ll become a habit.

Doing #30daysofpushup challenge. Day 1/30. Want to join me? Post your video. If you publicly commit, you’re more likely to see through the 30 days and then it’ll become a habit.

324,875 просмотров • 3 лет назад

The eagle has landed..

The eagle has landed..

147,595 просмотров • 1 год назад

Small language models give me warm, fuzzy feelings! Google dropped this 270M model (Gemma), and it’s absolutely wild how coherent and fast it is on my phone. Locally installed, chatting without internet (PocketPal app).

Small language models give me warm, fuzzy feelings! Google dropped this 270M model (Gemma), and it’s absolutely wild how coherent and fast it is on my phone. Locally installed, chatting without internet (PocketPal app).

60,437 просмотров • 1 год назад

I'm reiterating - I think a better interface for LLMs on desktop is an interactive notebook, not chat. I'm trying to build a better intuition for PCA and strongly feel an interactive notebook where I'm chatting with LLM + also modifying code/plotting figures is a much better interaction. Bonus points if you make notebook cells branchable (so if you get an error, you can ask LLM to fix it without adding that rotten context to the main branch of learning). Please steal this idea. Make an AI-powered, interactive, branching notebook.

I'm reiterating - I think a better interface for LLMs on desktop is an interactive notebook, not chat. I'm trying to build a better intuition for PCA and strongly feel an interactive notebook where I'm chatting with LLM + also modifying code/plotting figures is a much better interaction. Bonus points if you make notebook cells branchable (so if you get an error, you can ask LLM to fix it without adding that rotten context to the main branch of learning). Please steal this idea. Make an AI-powered, interactive, branching notebook.

58,232 просмотров • 1 год назад

I asked 6 agents in two conditions (parallel v/s sequential) to create the same tower defence game 👇 LLMs are known to have mode collapse, i.e. if you have multiple parallel runs on the same prompt, you get very similar responses. Is there a way to increase diversity? Yes, it turns out that if you do such multiple runs **sequentially**, instead of in parallel. (Plus for each sequential run, append results from previous runs and instruct the LLM to create something different from before.) I did this quick experiment to explore parallel v/s sequential runs for diversity. Given the same prompt to build a tower defence game, sequential runs had a much higher diversity and innovation v/s parallel runs. Check out this (AI-generated) video for gameplay clips. The sequential cohort created interesting new dynamics and visual style, while parallel ones all collapsed to similar colors and orbital style dynamics.

I asked 6 agents in two conditions (parallel v/s sequential) to create the same tower defence game 👇 LLMs are known to have mode collapse, i.e. if you have multiple parallel runs on the same prompt, you get very similar responses. Is there a way to increase diversity? Yes, it turns out that if you do such multiple runs sequentially, instead of in parallel. (Plus for each sequential run, append results from previous runs and instruct the LLM to create something different from before.) I did this quick experiment to explore parallel v/s sequential runs for diversity. Given the same prompt to build a tower defence game, sequential runs had a much higher diversity and innovation v/s parallel runs. Check out this (AI-generated) video for gameplay clips. The sequential cohort created interesting new dynamics and visual style, while parallel ones all collapsed to similar colors and orbital style dynamics.

11,581 просмотров • 1 месяц назад

Announcing a new project! Beautiful interactive explainers built using Claude Code in (almost) one shot. Launching with four topics: Fourier transformation, scaling laws in bio, cellular automata and LLMs. Link below 👇

Announcing a new project! Beautiful interactive explainers built using Claude Code in (almost) one shot. Launching with four topics: Fourier transformation, scaling laws in bio, cellular automata and LLMs. Link below 👇

25,829 просмотров • 8 месяцев назад

What if you have thousands of videos and want to search for clips within them? At Lossfunk, Shubham and Aaryan are building a video search engine. See how they search fight scenes within Varun Mayya's new @Project_11A launch video:

What if you have thousands of videos and want to search for clips within them? At Lossfunk, Shubham and Aaryan are building a video search engine. See how they search fight scenes within Varun Mayya's new @Project_11A launch video:

43,694 просмотров • 1 год назад

Google Search is now dominated by spam. So I made this Chrome extension that queries Reddit in the background, summarizes results and shows them in a sidebar next to google search results. It's like having your own Perplexity! Want it? Let me know!

Google Search is now dominated by spam. So I made this Chrome extension that queries Reddit in the background, summarizes results and shows them in a sidebar next to google search results. It's like having your own Perplexity! Want it? Let me know!

53,122 просмотров • 2 лет назад

Gemini 2.5 Pro is very good! Took me 15 mins to build this chrome extension that starts a new quiz on new tab to help remember countries, flags and their capitals. If you want it, let me know. I'll package it and release it on the store!

Gemini 2.5 Pro is very good! Took me 15 mins to build this chrome extension that starts a new quiz on new tab to help remember countries, flags and their capitals. If you want it, let me know. I'll package it and release it on the store!

31,218 просмотров • 1 год назад

How does Instagram decide which reels to show to you? How does it personalize your feed so fast? New explainer 👇

How does Instagram decide which reels to show to you? How does it personalize your feed so fast? New explainer 👇

16,889 просмотров • 7 месяцев назад

Are interactive essays the future of education? I asked Claude Code to build an interactive essay for PCA + SVD that works client side. Notice how python code executes and graphs are generated WITHOUT any backend! Play with it yourself 👇

Are interactive essays the future of education? I asked Claude Code to build an interactive essay for PCA + SVD that works client side. Notice how python code executes and graphs are generated WITHOUT any backend! Play with it yourself 👇

18,036 просмотров • 9 месяцев назад

🌀 Crazy project alert: What if you ask AI to visualize your state of mind? Presenting murmuration: a chrome extension that transforms your ChatGPT and Claude conversation topics into beautiful, animated black-and-white visualizations on every new tab.

🌀 Crazy project alert: What if you ask AI to visualize your state of mind? Presenting murmuration: a chrome extension that transforms your ChatGPT and Claude conversation topics into beautiful, animated black-and-white visualizations on every new tab.

15,178 просмотров • 7 месяцев назад