Chaitanya
@nutanc.bsky.social
CTO Co Founder of Ozonetel ($30 million ARR) in the CX space. Working on lots of open source AI projects.
The problem with current AI discourse is that you can just write science fiction as opinion. I like the podcasts by Dwarkesh, but in his new blogpost he is just stretching thin.
How many people on his podcast team? How many can be replaced by AI? Can he be replaced by AI?
Let's start there?
How many people on his podcast team? How many can be replaced by AI? Can he be replaced by AI?
Let's start there?
February 2, 2025 at 5:51 PM
The problem with current AI discourse is that you can just write science fiction as opinion. I like the podcasts by Dwarkesh, but in his new blogpost he is just stretching thin.
How many people on his podcast team? How many can be replaced by AI? Can he be replaced by AI?
Let's start there?
How many people on his podcast team? How many can be replaced by AI? Can he be replaced by AI?
Let's start there?
My suggestions for the 4 pillars of building Aai for India.
saasboomi.org/saas/ai/visi...
saasboomi.org/saas/ai/visi...
February 2, 2025 at 5:48 PM
My suggestions for the 4 pillars of building Aai for India.
saasboomi.org/saas/ai/visi...
saasboomi.org/saas/ai/visi...
Finally watches Manjummel Boys. Captures India perfectly.
1. No one follows rules.
2. If you are in trouble, don't count on the system to help you.
3. Your social net(friends and family) will always help you. That's what's great about India.
4. In the end, somehow, life moves on.
1. No one follows rules.
2. If you are in trouble, don't count on the system to help you.
3. Your social net(friends and family) will always help you. That's what's great about India.
4. In the end, somehow, life moves on.
January 15, 2025 at 2:49 PM
Finally watches Manjummel Boys. Captures India perfectly.
1. No one follows rules.
2. If you are in trouble, don't count on the system to help you.
3. Your social net(friends and family) will always help you. That's what's great about India.
4. In the end, somehow, life moves on.
1. No one follows rules.
2. If you are in trouble, don't count on the system to help you.
3. Your social net(friends and family) will always help you. That's what's great about India.
4. In the end, somehow, life moves on.
Thanks to social media future historians will have ample proof of how a civilization falls.
January 13, 2025 at 2:40 AM
Thanks to social media future historians will have ample proof of how a civilization falls.
Almost every edtech startup talks about changing the curriculum or making math fun or getting students ready for exams. Problem is, all of them become co curricular, that is, the student has to do something "extra" in addition to their school work.
January 12, 2025 at 3:22 AM
Almost every edtech startup talks about changing the curriculum or making math fun or getting students ready for exams. Problem is, all of them become co curricular, that is, the student has to do something "extra" in addition to their school work.
Just got the space in IIIT-H for the Viswam initiative. We also got funding for 8 full time researchers and 10 part time researchers. The space from where frontier research will happen and we will launch stuff which will help the world :) Will do a proper launch soon.
January 9, 2025 at 12:39 PM
Just got the space in IIIT-H for the Viswam initiative. We also got funding for 8 full time researchers and 10 part time researchers. The space from where frontier research will happen and we will launch stuff which will help the world :) Will do a proper launch soon.
Preparations for the AI conf. 1st day, a burning man style of event where people come together and build stuff using AI. All creators welcome.
We will also have round tables from the who's who of AI in India discussing future of AI.2nd day will be the conference with talks.Feb 15th-16th
We will also have round tables from the who's who of AI in India discussing future of AI.2nd day will be the conference with talks.Feb 15th-16th
January 7, 2025 at 4:45 AM
Preparations for the AI conf. 1st day, a burning man style of event where people come together and build stuff using AI. All creators welcome.
We will also have round tables from the who's who of AI in India discussing future of AI.2nd day will be the conference with talks.Feb 15th-16th
We will also have round tables from the who's who of AI in India discussing future of AI.2nd day will be the conference with talks.Feb 15th-16th
Most people have not read the ARC paper by @fchollet.
But because it's called ARC-AGI people are assuming we have achieved AGI. What has changed? No one tells.
So apparently we are not ready for what's coming.
Everyone will lose jobs.
Which jobs?
No one tells.
But because it's called ARC-AGI people are assuming we have achieved AGI. What has changed? No one tells.
So apparently we are not ready for what's coming.
Everyone will lose jobs.
Which jobs?
No one tells.
December 22, 2024 at 2:35 AM
Most people have not read the ARC paper by @fchollet.
But because it's called ARC-AGI people are assuming we have achieved AGI. What has changed? No one tells.
So apparently we are not ready for what's coming.
Everyone will lose jobs.
Which jobs?
No one tells.
But because it's called ARC-AGI people are assuming we have achieved AGI. What has changed? No one tells.
So apparently we are not ready for what's coming.
Everyone will lose jobs.
Which jobs?
No one tells.
I am saying we can say we have achieved AGI if an AI system can prove the Reimann Hypothesis.
December 21, 2024 at 7:47 PM
I am saying we can say we have achieved AGI if an AI system can prove the Reimann Hypothesis.
The reason the binary embeddings work so well is because the direction of the point(+/-) is as important as the distance of the point from the hyperplanes.
December 12, 2024 at 3:34 PM
The reason the binary embeddings work so well is because the direction of the point(+/-) is as important as the distance of the point from the hyperplanes.
India is a low trust country.Nobody trusts(with good reason) that you are doing a good thing for the right reasons
When I came back after my masters and told that I wanted to teach instead of doing an IT job,everyone assumed I came back because I didn't get a job in the US :)
When I came back after my masters and told that I wanted to teach instead of doing an IT job,everyone assumed I came back because I didn't get a job in the US :)
December 7, 2024 at 11:25 AM
India is a low trust country.Nobody trusts(with good reason) that you are doing a good thing for the right reasons
When I came back after my masters and told that I wanted to teach instead of doing an IT job,everyone assumed I came back because I didn't get a job in the US :)
When I came back after my masters and told that I wanted to teach instead of doing an IT job,everyone assumed I came back because I didn't get a job in the US :)
People say that AI will improve productivity and hence lead to abundance. Can people help me list out 5 cases where we can apply AI in tier 2 cities that can lead to productivity. Want to explore and see if we can identify 5 cases which we can guide 100,000 interns to go and implement.
December 6, 2024 at 2:51 PM
People say that AI will improve productivity and hence lead to abundance. Can people help me list out 5 cases where we can apply AI in tier 2 cities that can lead to productivity. Want to explore and see if we can identify 5 cases which we can guide 100,000 interns to go and implement.
It's important to use the right words when you are introducing new technology.
If a software does not work as expected, call it "hallucination".
That makes the model sound intelligent and that the problem is with you.
If makes it sound like your users don't know how to use it.
If a software does not work as expected, call it "hallucination".
That makes the model sound intelligent and that the problem is with you.
If makes it sound like your users don't know how to use it.
December 6, 2024 at 1:34 AM
It's important to use the right words when you are introducing new technology.
If a software does not work as expected, call it "hallucination".
That makes the model sound intelligent and that the problem is with you.
If makes it sound like your users don't know how to use it.
If a software does not work as expected, call it "hallucination".
That makes the model sound intelligent and that the problem is with you.
If makes it sound like your users don't know how to use it.
This is exactly the purpose of the 1 lakh interns yatra we are doing with Swecha and IIIT. Collect Telugu data with open licenses and train a foundational LLMs.
December 6, 2024 at 1:16 AM
This is exactly the purpose of the 1 lakh interns yatra we are doing with Swecha and IIIT. Collect Telugu data with open licenses and train a foundational LLMs.
Have to give it to Sama and the SV crowd for the way they innovate with pricing. Pro mode for a software. What does it do extra? It will "think more". 🤣🤑
I think we will also launch a new pricing tier for our software.
The "think more" tier.
I think we will also launch a new pricing tier for our software.
The "think more" tier.
December 6, 2024 at 1:15 AM
Have to give it to Sama and the SV crowd for the way they innovate with pricing. Pro mode for a software. What does it do extra? It will "think more". 🤣🤑
I think we will also launch a new pricing tier for our software.
The "think more" tier.
I think we will also launch a new pricing tier for our software.
The "think more" tier.
China $900 billion for education
India $13 billion
How can our next gen compete against this?
Solution?
I have no idea.But we can't give up. That's why this 1 lakh intern experiment is so important.
1 lakh*10,000(worth)=100 crore. Don't have budget.But have heart and volunteers.
India $13 billion
How can our next gen compete against this?
Solution?
I have no idea.But we can't give up. That's why this 1 lakh intern experiment is so important.
1 lakh*10,000(worth)=100 crore. Don't have budget.But have heart and volunteers.
Starting the 1 lakh internship program next week in Telangana.
Goals:
1. Train 1 lakh engineering students on AI.
2. Collect 1 billion tokens of Telugu.
3. Collect culture.
Looking for projects which students can do. What data does the AI community want us to work on?
Goals:
1. Train 1 lakh engineering students on AI.
2. Collect 1 billion tokens of Telugu.
3. Collect culture.
Looking for projects which students can do. What data does the AI community want us to work on?
December 2, 2024 at 3:24 AM
China $900 billion for education
India $13 billion
How can our next gen compete against this?
Solution?
I have no idea.But we can't give up. That's why this 1 lakh intern experiment is so important.
1 lakh*10,000(worth)=100 crore. Don't have budget.But have heart and volunteers.
India $13 billion
How can our next gen compete against this?
Solution?
I have no idea.But we can't give up. That's why this 1 lakh intern experiment is so important.
1 lakh*10,000(worth)=100 crore. Don't have budget.But have heart and volunteers.
Plans started for the AI conference in February. Goals are high this time. The goal is to be "the" AI conference for open source and the global south. Please suggest good speakers who are invested in open source AI and how AI can help the global south.
November 30, 2024 at 2:45 PM
Plans started for the AI conference in February. Goals are high this time. The goal is to be "the" AI conference for open source and the global south. Please suggest good speakers who are invested in open source AI and how AI can help the global south.
Starting the 1 lakh internship program next week in Telangana.
Goals:
1. Train 1 lakh engineering students on AI.
2. Collect 1 billion tokens of Telugu.
3. Collect culture.
Looking for projects which students can do. What data does the AI community want us to work on?
Goals:
1. Train 1 lakh engineering students on AI.
2. Collect 1 billion tokens of Telugu.
3. Collect culture.
Looking for projects which students can do. What data does the AI community want us to work on?
November 26, 2024 at 2:44 AM
Starting the 1 lakh internship program next week in Telangana.
Goals:
1. Train 1 lakh engineering students on AI.
2. Collect 1 billion tokens of Telugu.
3. Collect culture.
Looking for projects which students can do. What data does the AI community want us to work on?
Goals:
1. Train 1 lakh engineering students on AI.
2. Collect 1 billion tokens of Telugu.
3. Collect culture.
Looking for projects which students can do. What data does the AI community want us to work on?
AI is a solution waiting for a problem. Mostly.
MKBHD on Apple intelligence in summary: "some mildly amusing stuff that I will pretty much never use after this review". He was right about Humane pin, and he is right about that. www.youtube.com/watch?v=haDj...
An Honest Review of Apple Intelligence... So Far
YouTube video by Marques Brownlee
www.youtube.com
November 24, 2024 at 1:54 PM
AI is a solution waiting for a problem. Mostly.
When I started doing startups, like everyone else I thought building a great product and having a great pitch deck was paramount.
Boy was I wrong.
Everything is connections. Go to events. Make connections. Be visible. This is the age of hyperbole. Be loud.
Boy was I wrong.
Everything is connections. Go to events. Make connections. Be visible. This is the age of hyperbole. Be loud.
November 24, 2024 at 1:53 PM
When I started doing startups, like everyone else I thought building a great product and having a great pitch deck was paramount.
Boy was I wrong.
Everything is connections. Go to events. Make connections. Be visible. This is the age of hyperbole. Be loud.
Boy was I wrong.
Everything is connections. Go to events. Make connections. Be visible. This is the age of hyperbole. Be loud.
If you break it down, AI is actually very simple.
We already have huge insights. Just need some time to crack this open.
We already have huge insights. Just need some time to crack this open.
November 24, 2024 at 1:24 PM
If you break it down, AI is actually very simple.
We already have huge insights. Just need some time to crack this open.
We already have huge insights. Just need some time to crack this open.
It's really surprising how in AI the best things are the easiest ones.
I mean many times I believe, kNN is all you need :)
I mean many times I believe, kNN is all you need :)
November 23, 2024 at 6:42 AM
It's really surprising how in AI the best things are the easiest ones.
I mean many times I believe, kNN is all you need :)
I mean many times I believe, kNN is all you need :)
The 1 lakh AI internships with students will be starting next month.
This will be the biggest open source AI data collection effort in the world. Target is to collect 1 billion Telugu tokens.
This will be the biggest open source AI data collection effort in the world. Target is to collect 1 billion Telugu tokens.
November 23, 2024 at 2:59 AM
The 1 lakh AI internships with students will be starting next month.
This will be the biggest open source AI data collection effort in the world. Target is to collect 1 billion Telugu tokens.
This will be the biggest open source AI data collection effort in the world. Target is to collect 1 billion Telugu tokens.
Turns out, you can use our shape of stories approach to analyse web site content
gpt3experiments.substack.com/p/using-sema...
gpt3experiments.substack.com/p/using-sema...
Using shape of stories for web site analysis
Can we use information from semantic chunks to analyze website content
gpt3experiments.substack.com
November 20, 2024 at 6:52 PM
Turns out, you can use our shape of stories approach to analyse web site content
gpt3experiments.substack.com/p/using-sema...
gpt3experiments.substack.com/p/using-sema...
List of AI activities I am planning for 2025
November 20, 2024 at 3:27 AM
List of AI activities I am planning for 2025