By Akhilesh Tilotia
A trial digest for the week from Sun, Nov 1 to Sat, Nov 7. Two items are summarised. One audio episode (Ajay Shah on Decoded) has no published transcript, so I made a machine transcript and read it in full except for short garbled stretches (about one minute near 41:08, and short stretches near 32:42 and 5:00). A written newsletter post is the second. One more audio episode could not be transcribed well enough and is listed at the end, not summarised.
India
Decoded: Ajay Shah on institution building, AI, and the reform India needs
Guest: Ajay Shah. Hosts: Shrreya Loomba and Aly Muhammad Dadwani. Released Wed, Nov 4.
Ajay Shah is an economist and co-founder of the XKDR Forum, a Mumbai non-profit research group. He trained as an aeronautical engineer at IIT Bombay, worked in academia, policy research and the Ministry of Finance, and co-hosted the podcast Everything is Everything with Amit Varma. This is the first episode for the hosts, Shrreya Loomba and Aly Muhammad Dadwani, as co-hosts.
- Institutions, not clever people, make places like Bell Labs, RAND and the Cavendish Laboratory work. He and Susan Thomas built research teams in India on that idea, and he says small, artisanal groups are what they know how to run.
- He names three great questions of the age: defending economic growth, climate change, and populism. His group works on the first, tries to help on the second, and treats the third as beyond its reach. He says a general toolkit of market failure and state capability lets a team be useful across many fields.
- On AI and learning, he says homework and projects no longer work as education. Handing intellectual work to an LLM costs you your own growth. Two uses he backs: AI as a tutor that sets reading lists and diagnostic exams, and AI as a critic that finds mistakes in work you did yourself.
- Drawing on Terence Tao, he says a proof nobody understands is not mathematical knowledge. He would not give publication status to an AI proof until a human can explain it, and says incentives like prizes and tenure will need rethinking.
- His one reform for long-term growth is removing barriers on trade, capital, services and people, and treating foreign and Indian firms alike.
Climate and energy
Hannah Ritchie, By the Numbers: The IEA needs to make its data open
Author: Hannah Ritchie. A written newsletter post, not an interview. Released Tue, Nov 3.
Hannah Ritchie works at Our World in Data, which she calls her main job, and writes the By the Numbers newsletter. She is the author of the book Clearing the Air. The post argues that the International Energy Agency should make its data free to use.
- The IEA is mostly funded by taxpayers in its 32 member countries, yet users pay thousands of euros for a license and may not share the data or any analysis that relies on it.
- Free sources do not fill the gap. Our World in Data uses Ember for electricity and the Energy Institute for primary energy, but there is no free data on final energy and coverage in lower-income countries is thin.
- She lists four costs: lower-income researchers are locked out, researchers waste time on workarounds, results cannot be replicated, and public debate suffers in a period of price shocks.
- Her fix is a donor. She estimates IEA sales at roughly 6 to 10 million euros a year, so five years of free data would cost about 30 million euros, and she suggests very wealthy people, for example from tech listings, could fund it.
Not read in full, so not summarised:
– Puliyabaazi, “How can states improve law and order” (Thu, Nov 5): episode page. Hindi and English audio with no published transcript. I tried a machine transcript, but the Hindi came out too garbled to rely on. The show notes list police and judicial capacity, state competition, and police reform.
Nothing in the window: Conversations with Tyler (latest episode Thu, Oct 15), the Ajay Shah newsletter (latest post Tue, Oct 27), and Ideas of India (no Nov 5 episode could be confirmed, since its page lists Sep 24 as the latest). Puliyabaazi’s Wed, Nov 4 post is a magazine issue, not an episode. Our World in Data’s Nov 2 data insight on mental health was not covered.
Built by Akhilesh Tilotia, with assistance from Instinct.
