Abhivardhan

82.6K posts

Abhivardhan banner
Abhivardhan

Abhivardhan

@IndusThink

AI Governance @IndicPacific. President @IndianSocietyAI. Posts personal, unrelated to organisations I am affiliated with. My #AI book: https://t.co/IdknPNfVVL

Indo-Pacific ⇄ Europe ⇄ World Katılım Haziran 2015
1.6K Takip Edilen3.1K Takipçiler
Sabitlenmiş Tweet
Abhivardhan
Abhivardhan@IndusThink·
June has been unexpectedly a very grateful month. Third article on AI and geopolitics :) Happy that I pursued an idea that might have its way: India needs an AI dimension in its Act-East Policy. Glad to have co-proposed some ideas on this industry policy pivot with @AVIRAL96 Read on @Diplomat_APAC: thediplomat.com/2026/06/the-ai… cc @ThisIsSoliman @adityajakki @harshmadhusudan @RupakChatto @RMantri @d_extrovert @jppjagannath1 @GenevieveDonn
English
3
7
26
4.4K
Abhivardhan
Abhivardhan@IndusThink·
@LegalTalwar I don't wanna talk on the state of opportunism in legal journalism anyways. LawBeat is good. Didn't feel too parochial to me to say the least. Just saying what I myself noticed. Not commenting on others.
English
1
0
4
1.8K
Sanya Talwar
Sanya Talwar@LegalTalwar·
Genuine question. “Glad I left LawBeat for good” because it invited Abhijit. Fast forward a few weeks: smiling through a national TV debate with the very same person. How come was there no hesitation in joining him on NDTV? Perhaps because she has realised she needs the visibility alongside you on big media or she genuinely likes you now @Iyervval ?
Sanya Talwar tweet mediaSanya Talwar tweet media
English
18
79
579
32.6K
Abhivardhan
Abhivardhan@IndusThink·
Don't get me started on how MLAs behave. Also don't get me started how judicial polyvocality that can possibly happen in High Courts. Sorry man but now it seems you are also parading for an ecosystem. Main to bhai hoon equal opportunity offender. LW hates me, RW doesn't understand me.
English
0
0
0
50
कुशल मेहरा
If any student gets an FIR for merely criticising the Prime Minister or abusing him profusely, then I do not support that. I feel everybody, including the PM, can be abused. It should not be illegal. You can debate the morality based on your social etiquette, but make it legal. It does not mean people need to like it. Also, if you really want to get a reality check, try abusing your local MLA, MP, or corporator in India. Now that is the real test. THANK YOU FOR YOUR ATTENTION TO THIS MATTER
English
90
103
918
37.6K
Abhivardhan
Abhivardhan@IndusThink·
@avataram But intelligence is not abundant due to AI still. What you are referring to is conviction and courage. That I agree with.
English
0
0
2
104
Abhivardhan
Abhivardhan@IndusThink·
No @amargov saar - tech sovereignty is not an overused word. These Delhi-based policy affairs professionals are mocking the Indian sovereign AI ecosystem necessity right now. We need tech sovereignty. When we build meaningful things, we create material prosperity. I am with you.
Amar Govindarajan@amargov

I'll use this opportunity to put out some info on the direction change at @SwarajyaMag, our perspective on what we should be doing and why I've pushed the org into these areas. I do think tech sovereignty + material prosperity will be the key driver of an Indian renaissance. It maybe an over used word but let it be. (Other main drivers will be culture and governance) It takes a cultural/mindset shift to consciously work towards tech sovereignty. Our job is to articulate this and the challenges in areas of talent, know-how and building out the industrial muscle required. Of-course we also celebrate the wins. While doing the above we're trying to not look at everything through the lens of govt - there is more than govt to all this - industry, media, overall society's outlook towards hard tech/trade work and manufacturing. Once you snap out of this false belief that every thing you do has to be centered around govt failures/successes you can actually focus on the story - and that works out to be more fair to govt/industry. In this particular instance detailing out the immense challenges in semiconductor space deflates shallow criticism about 'govt failure' - how can govt get all these impossible things done in 5-6yrs? The industrial muscle will have to be built over decades and we could discuss policy support necessary to make it happen. It's a better idea to look at us as an intellectual project that speaks to a kind of Indian who understands what an Indian renaissance means in its entirety (industry, tech, culture, governance, society)

English
0
1
5
153
Abhivardhan
Abhivardhan@IndusThink·
@kushal_mehra Maine blink kiya pehle ki aap sambodhit kisko kiye bhaisaab🤣 Got it after a few seconds.
HT
0
0
0
113
Manasa Manjunath
Manasa Manjunath@ManeeManjunath·
We do not talk about ethics enough when we speak about AI. The primary concern is job loss but not how the job loss is coming about. Every AI company has unethically gathered data to train its models. There has been no pushback on this front from both governments & the public.
Barbarian Herodotus@BBHerodotus

After seeing a post on it yesterday, I’ve looked more deeply into Anthropic’s purchasing and destruction of millions of used books. They spent tens of millions of dollars acquiring and destroying books to train their AI model, Claude. They knew, from the beginning that it would be unpopular. They knew what they were doing, was wrong. An internal memo from 2024 stated: “Project Panama is our effort to destructively scan all the books in the world.” “We don’t want it to be known that we are working on this.” Why used books? Companies like Anthropic and Meta found it impractical to gain permission from authors and publishers to use their work. So instead, they stole their work to train their product through mass purchases of used books and through pirating them from the internet. The authors of these works, didn’t receive a dime for their efforts. These companies stole the hard won literary talent of others to train machines to replace the victims of their theft. Anthropic cofounder Ben Mann personally downloaded countless books for this purpose from LibGen, a book pirating website. When a new book piracy website debuted in 2022, Mann sent an email sharing the link with his employees stating: “just in time!!!” Meta employees were wary about the piracy of books, with one engineer writing, “Torrenting from a corporate laptop doesn’t feel right.” Anthropic purchased millions of books, in batches of tens of thousands from Better World Books and U.K.-based World of Books. They then used a “hydraulic powered cutting machine” to slice off the spines of these books before scanning them. Finally, the remains of these books were picked up for recycling. Feel free to use AI, it’s a free country. But remember that, if you use Claude, that Anthropic has committed a great wrong, and they knew that they were doing it from the beginning. Below is an image, from a court case, of a warehouse full of used books, that instead of being held and read by you and your children, have been destroyed forever. AI is inevitable, but spend your money on the companies who do good, and do not knowingly commit evils. May the most moral company win.

English
1
3
14
810
Abhivardhan
Abhivardhan@IndusThink·
@ManeeManjunath There is a funding pattern here actually. E-acc etc. and so on. Ancillary risks are more problematic and connected to real life.
English
1
0
1
17
Manasa Manjunath
Manasa Manjunath@ManeeManjunath·
@IndusThink I guess it’s also because extremes get more interest? I’ve hardly heard talk of ancillary risks. Will need to read up about it.
English
1
0
1
18
Amitabha Bandyopadhy
Amitabha Bandyopadhy@abandopa·
Wow!!! Such an interesting story @Fintech03. Thanks for sharing. These stellar personalities succeeded despite the system.
Parimal@Fintech03

Picture a roaring, icy river slicing through the isolated Spiti Valley in 1972. The local villagers are completely cut off by a mountain collapse. The military says a traditional concrete/riveted steel bridge will take yrs to bring heavy machinery up the treacherous cliffs. Then, a young woman steps out onto the frozen mud wearing a hard hat, carrying a set of rolled up blueprints under her arm. She does not have bulldozers. She does not have massive cranes. Instead, she brings trucks loaded with small, lightweight steel components, like a adult-sized structural Lego set. In just 4 months, w/o a single piece of heavy machinery, her team hand-assembles a rock-solid, load-bearing steel bridge across the gorge. Her name was Shakuntala Bhagat & she was changing the rules of civil engineering forever. Born in 1933, Shakuntala was the daughter of S.B. Joshi, a legendary figure known as the "Father of Bridge Engineering in India". While other young girls of the 1940s were expected to follow quiet, traditional paths, Shakuntala spent her childhood staring at bridge trusses & stress diagrams. In 1949, she walked into the classrooms of VJTI in Bombay. The lecture hall went dead silent. Dozens of male students turned to stare at the only woman in the room. In 1953, at just 20 yrs old, she walked out with a degree, making history as India’s 1st female civil engineer. She flew to the United States, earned her Master's in Structural Engineering from Penn & trained across Germany & the UK. When she returned home in the 1960s, she joined IIT Bombay as an Assistant Prof, taking charge as the Head of the Heavy Structures Lab. At IIT Bombay, Prof. Shakuntala was obsessed with a single problem: Why did Indian infra take so long to build?Traditional steel bridges relied on slow, dangerous, on-site welding & heavy riveting. If a flood wiped out a road in Arunachal Pradesh/Kashmir, cut-off communities had to wait yrs for a replacement bridge. In the late 1960s, turning her small Bombay apartment into a late night workshop littered with steel prototypes & stress formulas, Shakuntala & her engineer husband, Anirudh Bhagat, cracked the code. They invented & patented the Quadricon Modular Bridge System & the Bhagat Unishear Connector. Instead of custom-casting a bridge for every river, their system used mass-produced, interchangeable, lightweight steel modules. They could be packed onto small trucks, hauled up narrow mountain tracks & assembled on-site using simple hand tools, reducing bridge construction time from yrs to weeks. Yet, when they took this revolutionary idea to institutional investors & govt departments for R&D funding, they were met with cold skepticism. "If advanced Western nations have not developed a total-system modular steel bridge like this, how can an Indian couple do it?" The establishment refused to back them. She & her husband took massive personal financial risks, mortgaging their own assets to build & test the prototypes themselves. In 1972, the Invention Promotion Board awarded them the highest honor for the Unishear connector. The skeptics were forced to eat their words. Shahkuntala Bhagat went on to design & oversee the construction of 200+ Quadricon steel bridges worldwide, including 69 vital superstructures across India’s most dangerous border terrains, from the high passes of Kashmir to the dense forests of Arunachal Pradesh, as well as projects in the US, the UK & Germany. In 1993, she was named Woman Engineer of the Year. When she passed away in 2012 at the age of 79, she left behind an army of students & a country connected by her steel. We celebrate the famous monuments of concrete and glass in our metros, yet few history books mention the woman in a hard hat who stood on freezing riverbeds to connect India's most forgotten corners. The next time you travel through a remote mountain pass in the Himalayas & your vehicle rolls smoothly over a steel bridge spanning a deadly gorge, look closely at the steel joints beneath you. You are driving across the indestructible genius of Shakuntala Bhagat.

English
1
0
4
387
Abhivardhan
Abhivardhan@IndusThink·
See - it gets saturated when you get to extremes. I am a middle ground guy and try people to enable more real and nuanced angles. I have known people in catastrophic risk institute-like places too. Ancillary risks need review and some interventions. Even non-ancillary ones have some limited merit. But claiming first that an LLM/agent is replacing people and then conflating what is an LLM and what is an agent sounds baffling to me. AI 2027 was overestimated as well by design no matter whatever innocence they claim.
English
1
0
0
16
Manasa Manjunath
Manasa Manjunath@ManeeManjunath·
@IndusThink Just from listening to the podcasts, I felt there was a tendency towards both extremes & like you say, proving their theory right. I can think of only a few nuanced ones who treaded a ‘middle path’. Do you really think it’s saturated? I think they were very limited approaches.
English
1
0
1
23
Abhivardhan
Abhivardhan@IndusThink·
It is more indicative of lesser work because of a saturated estimate of how AI safety works. People overhyped on Ai safety issues by focusing on some issues rather irresponsibly, including those of METR and AI 2027. It sometimes felt they wanna prove themselves “right” or they want something bad to happen. But even cautioning should not be fearmongering. 🥲 That is why Anthropic is facing backlash since Jan 2026.
English
1
0
0
15
Manasa Manjunath
Manasa Manjunath@ManeeManjunath·
@IndusThink I mostly follow the developments on podcasts so I wasn’t aware of this work. I’m assuming the concentration on LLMs is because they’re currently the ones in use & at risk of abuse. I’ve noticed that there has been a drop off on podcasts on safety, alignment & ethics.
English
1
0
1
22
Abhivardhan
Abhivardhan@IndusThink·
A friend has triggered me so I am saying this. I am very fast in typing🤣. After publishing 64 downloadable chapters, papers, books and white papers and then 20+ blogs/write ups and media features in 8 years, obviously bhaijaan I type like Usain Bolt🥹😂😂😂😅. Does it mean I am sitting free? Please make your dumb judgment on me. I have no time explaining how busy, stressed and at times confused I am. You have no idea folks. Bas now don’t force me to say more. Jiyo aur jeeno do na. 🥹 Actually I know why some people felt by mistake that I am “rude”🤣 1. English is a rude language. Yahi Hindi me likha hota to rudeness wala stress nahi aata (@adityajakki phupha knows this) 2. ⁠Second, Whatsapp is a chatting app. Hence the notifications feature + Chatting feature UI makes it overwhelming for people. This is where I sympathise with people. But yaar I am not your enemy nor am I rude. We are all stressed by notifications and overuse of smart phones. 🤣🥹 Ab theek? Mereko baksh do. cc @AKLKO1977
English
0
1
0
105