Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Deedy Profile picture
Deedy
@deedydas
VC at @MenloVentures. Formerly founding team @glean, @Google Search. @Cornell CS. Tweets about tech, immigration, India, fitness and search.
8 subscribers
Mar 28 • 17 tweets • 5 min read
Compilation of 16 USEFUL things people did with GPT 4o native image generation can do (aside from Ghibli images) in the last 48hrs!

They include
— Full video movies / podcasts / ads
— UI design loop
— 3D model gen

Full Lord of the Rings in Ghibli

1/16 Videos in different animation styles

2/16
Mar 15 • 11 tweets • 3 min read
Google's new AI image editing model is going to kill 99% of Photoshop.

By just describing your image edit in English, you can:
— get passport picture from photo
— make me look like I'm on my way to work
— decorate this house for me
— put these clothes on me
— make a map in a game from these sprites
— animate this sprite
— combine all these faces together
— show this model holding my product
— make me bald / give me hair

You can access Gemini Flash 2.0 experimental on Google AI studio.

This is a 1B+ user product waiting to happen. 1. make me look like I'm on my way to work
Mar 1 • 4 tweets • 1 min read
BREAKING DeepSeek just let the world know they make $200M/yr at 500%+ profit margin.

Revenue (/day): $562k
Cost (/day): $87k
Revenue (/yr): ~$205M

This is all while charging $2.19/M tokens on R1, ~25x less than OpenAI o1.

If this was in the US, this would be a >$10B company. Image Source: github.com/deepseek-ai/op…
Feb 28 • 10 tweets • 3 min read
Can confirm. GPT 4.5 is hilarious.

"Finish the text:
> Be me
> L3 at Meta / Google
>" Image Okay I decided to do something that turned out better than I thought. Ran the same prompt for L4-L12.

For those who don't know, you start at L3, L5 is Sr, L6 is Staff / Manager, L7 is Senior staff, L8 is director, L9 is senior director, L10 is VP, L11 is SVP, L12 is CEO

L4 Image
Feb 25 • 4 tweets • 1 min read
If you haven’t used Grok 3 “Sexy” 18+ mode, drop everything and try it.

I can’t explain how unbelievably messed up this is (and I can’t post a video). This may single handedly bring down global birth rates.

I can’t believe Grok actually shipped this. Image I should mention this was Grok voice mode, so there was no prompting involved.

The other issue is.. it sometimes completely loses context of what gender it is you and you are 😆
Feb 20 • 8 tweets • 2 min read
The last 24 hours has had SO much tech news.

— Microsoft Majorana 1 quantum chip
— Google AI co-scientist
— xAI Grok 3 model release
— Clone musculoskeleton
— Claude reasoning coming soon
— Arc institute biological model
— Sakana AI CUDA kernel optimizer

All the threads here: Microsoft
Feb 6 • 4 tweets • 2 min read
PDF parsing is pretty much solved at scale now.

Gemini 2 Flash's $0.40/M tokens and 1M token context means you can now parse 6000 long PDFs at near perfect quality for $1 Image
Image Source: sergey.fyi/articles/gemin…
Jan 21 • 4 tweets • 1 min read
HUGE Immigration News! The birthright citizenship executive order Trump signed today means children of non-citizens or non-permanent residents will not be US citizens.

Children of Indians who are in the 100+yr green card wait WILL NOT be US citizens.

Applies from Feb 20, 2025. Image Source: whitehouse.gov/presidential-a…
Jan 20 • 6 tweets • 2 min read
HUGE: DeepSeek R1 from China reaches o1 level performance! Most important paper of 2025.

They show that using pure RL using GRPO can teach models to reason better and better with more tokens: the first at this scale.

Here are the 4 top research learnings from the paper:

1/5 Image 1. Distilling more powerful models like R1 on smaller base models (Qwen / Llama) outperforms large scale RL training on those base models themselves (like QwQ-32B).

2/5 Image
Jan 4 • 12 tweets • 4 min read
MEGATHREAD: EB-1s for Indians in the US 2025!

— It's been 2yrs since the priority date has moved past Feb '22
— Indian EB-1 petitions stayed flat this year, at ~12k
— The backlog moved forward by ~200/mo.
The EB-1 backlog is basically stuck.

Let's break it down..

1/12 Image What is the EB-1?

Indians under the standard EB-2/3 green card category have a 100+yr long wait. This is because of the 7% country cap by category.

This was the initial misconstrued old remark by Sriram that was taken out of context and set off this whole H-1B drama.

2/12
Jan 3 • 5 tweets • 3 min read
The Superhuman AI for Poker is a timeless research paper.

Poker, unlike Go/Chess, is an imperfect info game. Pluribus, made by now OpenAI engineer, beat 5 pros decisively in 10,000 poker hands, winning 48 milli big blinds/game.

Made a mini demo on rock-paper-scissors too:

1/5 Image Like humans, Pluribus confirms that limping, calling the “big blind” rather than folding or raising, is suboptimal.

However it disagrees with the wisdom that “donk betting” (starting a round by betting when one ended the previous betting round with a call) is a mistake.

2/5 Image
Jan 2 • 11 tweets • 2 min read
BOMBSHELL: New Hindenburg report calls Carvana "a father-son accounting grift for the ages."

The online car dealer's stock went 60x in the last 2yrs after dropping 99%, hitting a $44B valuation, while the CEO's father sold $5B in stock.

Here's the wild story of $CVNA...

1/10 Image From Aug 2021 to Dec 2022, Carvana stock plunged 99%. CEO's father Ernest Garcia II sold $3.6B in stock before that.

After the 63x rebound, he just sold another $1.4B.

Funny timing?

2/10
Dec 30, 2024 • 8 tweets • 3 min read
H-1B DATA! SALARY VS INDIAN % show the 2 sides of the H-1B

~380K accepted H-1Bs (not LCAs) from 2021-24. We plot the top 200 employers by volume: avg base salary vs %age of H-1Bs that were Indian.

The entire H-1B is NOT low wage Indian. ~30-50% is.

Let's unpack the data🧵

1/7 Image The cluster of large employers that pay <$100k and 95%+ of their applicants are Indian are the IT staffing firms

Infosys
TCS
Cognizant
Wipro
IBM
Capgemini
Accenture
L&T
HCL
Tech Mahindra

Including the long tail of smaller firms shows ~50% of H-1Bs are in this bucket.

2/7 Image
Dec 29, 2024 • 5 tweets • 2 min read
We can't argue about the H-1B without data, so here it is:
Top 200 H-1B Employers by Salary from 2021-2024!

The red bars are companies with 95%+ Indians — the IT staffing firms, all in the bottom 50%. Bottom 70 are <$100k.

~30-50% H-1B has a low-wage problem. Image Source: github.com/BloombergGraph…
Dec 23, 2024 • 13 tweets • 2 min read
Starlink is one of the seminal feats of engineering in history.

It will enable internet that's
— fast 100-300mbps
— uncensored
— cheap $1500/yr
in:
— the most remote areas
— ships in the ocean
— airplanes in the sky
— poles

But few even know what this picture is..

1/10 Image Traditional satellite internet uses geostationary orbit (GEO) - satellites at 36,000km altitude. The physics is simple but the latency is brutal: 600ms+ for signals to make the round trip.

Online gaming? Video calls? Forget it.

2/10
Dec 22, 2024 • 16 tweets • 10 min read
OpenAI o3 scored 91.5% on public ARC-AGI test despite taking 57M tokens and ~$5000/task. The problems are fairly simple!

I made an app with Claude to visualize them. Strangely, Sonnet can even identify the pattern sometimes.

Let's look at the 34 problems o3 got wrong...

1/13 Image The spiral with a margin of one.

2/13 Image
Image
Image
Dec 19, 2024 • 8 tweets • 2 min read
HUGE immigration news on the H-1B!

— Cap-exempt H-1B (no lottery) applies to more companies
— Founders can self-petition and own >50%
— Easier hiring for startups
— F-1 students get cap-gap until April 1
— Crack down on IT consulting fraud

Let's get into it 🧵

1/7 Image More orgs are cap-exempt employers.

Beyond the 85k H-1B limit, certain employers could bypass the lottery. Now the definition changes from "primarily engaged" in research to it being a "fundamental activity", unlocking more companies. Part-time work is allowed.

2/7
Dec 17, 2024 • 5 tweets • 2 min read
o1-preview is far superior to doctors on reasoning tasks and it's not even close, according to OpenAI's latest paper.

AI does ~80% vs ~30% on the 143 hard NEJM CPC diagnoses.

It's dangerous now to trust your doctor and NOT consult an AI model.

Here are some actual tasks:

1/5 Image Here's an example case looking at phosphate wasting and elevated FGF23, then proceeded to imaging to localize a potential tumor.

o1-preview suggested testing plan takes a broader, more methodical approach, systematically ruling out other causes of hypophosphatemia.

2/5 Image
Dec 16, 2024 • 10 tweets • 4 min read
Want to design better AI agents? Take notes from code writing systems.

Techniques include
— Multi-agent
— Tool choice
— Underlying model
— Diff format
— Innovative Signals
— Code retrieval + knowledge graphs
— LSP
— Fault localization

Let's dive deeper with real examples:

1/10 Multi-agent

Use agents with different roles / prompts that have access to different tools and can hand off to another agent.

Some roles used in coding: searcher, planner, reproducer, coder, tester, editor

2/10
Dec 10, 2024 • 5 tweets • 2 min read
As cool as the new Sora is, gymnastics is still very much the Turing test for AI video.

1/4 Attempt 2.

2/4
Dec 10, 2024 • 6 tweets • 3 min read
HUGE Immigration News!

We have the first EVER look at H-1B lottery data. Did you also suspect the lottery wasn't truly random? They're not.

Certain companies like Tiktok and Bytedance have 50% higher odds than average.

I broke it down by nationality, company and age...

1/5 Image
Image By Age.

In 2024, there were 350,084 applications and 85,304 were selected, a 24.4% acceptance rate.

The process seems to have rampant ageism, with only ages 26-32 having above average acceptance rates.

A 26yo has 50% higher odds than a 36yo+!

2/5 Image