Joined June 2019
47 Photos and videos
This is another way to define " doing what you love"
22 May 2025
Your first task is to find what feels effortless to you. Your second task is to put maximum effort into it.
25
But before this season ends, I want Mercedes.
If FP2 in Bahrain GP means anything then Max va Fernando is going to be one hell of a season: both 2x WDC, Max literally does not give 2 cents and Fernando has a lot of chips on his shoulder.
2
45
If FP2 in Bahrain GP means anything then Max va Fernando is going to be one hell of a season: both 2x WDC, Max literally does not give 2 cents and Fernando has a lot of chips on his shoulder.
96
RandomTrees retweeted
18 Feb 2023
Imagine having 4 cofounders at a VC funded startup 🥹 Selling for $100M Diluted to 10% ownership by VCs Tax takes 50% Go home with $1.25M cash after 6 years working 7 days a week 12 hours per day or $47/hour Many people on Upwork make more than that!
455
304
4,812
2,542,127
RandomTrees retweeted
17 Feb 2023
Here's a GPT implementation with NumPy in 60 lines of code:
62
525
3,483
549,401
RandomTrees retweeted
Classic example of Indian dhandha - we create ways to make a quick buck (taken from the book 'CEO Factory')
23
94
1,052
98,721
RandomTrees retweeted
How useful is collecting & analysing tons of data is a question I ask my peers. We have no data team at @zerodha as we believe it doesn't give useful insights & also since we never upsell. Article & comments on this topic 👇are interesting news.ycombinator.com/item?id… This 😬
169
353
3,768
552,843
RandomTrees retweeted
Unlock blazing fast GPU inference for your generative models! Our team achieved 2.3x improvement on @OpenAI Whisper (large) vs HuggingFace FP16 PyTorch 2.0. CUDA Graphs & a simple trick for avoiding OOM, follow this thread to find out more (links at the end) 💥
9
56
390
73,198
RandomTrees retweeted
I think Yann is (definitely) right. But I also think even LLMs are more useful than cats. This is the whole crux of my ambivalence to AGI debates.
5 Feb 2023
Before we reach Human-Level AI (HLAI), we will have to reach Cat-Level & Dog-Level AI. We are nowhere near that. We are still missing something big. LLM's linguistic abilities notwithstanding. A house cat has way more common sense and understanding of the world than any LLM.
2
2
9
37,904
This @ylecun guy
so are we all saying HLAI instead of AGI now?
8
That's exactly what I said a couple of days ago. @ylecun is the new @GaryMarcus
I am not the only one who has noticed what’s up here.
11
Makes sense to me
👀 @Sama (CEO of @OpenAI) is no longer following @ylecun
29
RandomTrees retweeted
29 Jan 2023
If you are building in public, what do you love most about it?
99
7
141
40,563
RandomTrees retweeted
Replying to @ylecun
When DCNNs said 'cat' in response to an image, or when VAEs drew an image of cat they met with all but praise. Nobody said 'they make stuff up'. ChatGPT is a milestone in AI and big cos should swallow their pride and start working, otherwise they will be sidelined.
4
1
27
2,848
1) @ylecun 's behaviour is anti- AI and for-me atm. Alarmingly unprofessional. Just like @SchmidhuberAI or @GaryMarcus . Just sad 2) feeling bad for the Google LaMDA test Engineer he was on point and tells you all big tech have the same stuff :) #happy
29 Jan 2023
Nobody, including the most knowledgeable ML and NLP people, believed that a chatbot as capable as ChatGPT could be possible anytime soon, especially as a language model. It looks embarrassing when some of those people now try to diminish the importance of OpenAI achievement.
19
If it is asking for money, then it's a liability. If it's generating money, it's an asset. Samay samay ki baat hai.
Is house an asset or a liability?
11
All the holders of XGBOOST coins ft @tunguz
32
RandomTrees retweeted
ChatGPT is just mass scale user testing in final stages of GPT-4 prep right?
63
67
1,914
RandomTrees retweeted
I'm 30 years old. And I run a $200M portfolio of businesses. This is my daily routine:
2,772
1,004
10,110
It's great that @deepakshenoy called out @Akshat_World . In a lot of his videos, Akshat has the same pattern of argument: 1) some vague trend or data that is easy to spot/understand 2) that trend/data gets connected to whatever he is trying to sell atm. 1/
This is wrong - govt debt is not even 1x GDP even now. Even all the money in the country - printed and in bank accounts - barely adds up to 1x GDP Sure, invest in equity, but have the right reasons :)
1
In his last video, he recommended to buy "tech stocks*!! - the market is at an all time high - the "tech" companies are losing a lot of value (have already lost, will lose some and more) - tech valuations are VC valuations and aren't worth a sachet of shampoo ;)