In March 2025, Dario Amodei, the CEO of Anthropic, predicted that 90% of code would be written by AI as early as June 2025 and no later than September 2025. This turned out to be dead wrong. 

Amodei claims that 90% of code at Anthropic (and some companies they work with) is being written by AI

There are some interesting comments on LessWrong

I have been recommended the Drone Ultimatum podcast, but haven't listened to any

Discussed on the forum here

I am a huge Ted Chiang fan, but your review misses one of the most amazing things about his writing -- it is written in very brief and straightforward words and sentences!

A number of his short stories are available for free online, for instance Exhalation in Lightspeed magazine. Under 'Works' on his Wikipedia page, you can find others (sometimes via web archives)

A large recent RCT found that free contraception had no impact on birth rates in Burkina Faso - I wonder if/how this affects this cause area

One website, which doesn't quite match all your criteria, is https://longbets.org/bets/, which has bets from people like Warren Buffet, Scott Alexander, Ray Kurzweil, and Eric Schmidt

GiveWell's cost to save a life has gone from $4,500 to a range between $3,000 and $5,500:

https://www.givewell.org/how-much-does-it-cost-to-save-a-life

From at least as early as December 2023 (possibly as early as December 2021 when the page says it was first published) until February 2024, that page highlighted a $7.2 million 2020 grant to the Against Malaria Foundation at an estimated cost per life saved of $4,500.

The page now highlights a $6.4 million 2023 grant to the Malaria Consortium at an estimated cost per life saved of $3,000.

You can see all the estimated cost per life saved (or other relevant outcome) for all GiveWell's grants at this spreadsheet, linked-to from:

https://www.givewell.org/impact-estimates

Many such cases

Luckily those suggestions are all useful for SEO too!

Some other things to consider (from figures like Tyler Cowen^[1] and Patrick McKenzie^[2] (edit: also Gwern) who talk about how their primary audience is now LLMs):

• Prefer short paragraphs that restate names, titles, other nouns (rather than using pronouns ("he"/"she"/"it")
  • So that LLMs can lift paragraphs wholesale, and minimises errors
• Consider what license your writing is published under
  • A more permissive license increases the likelihood of being included in training corpuses
• Design headlines to mirror exactly what a user's prompt might be
• Modular layouts – Bullets, numbered lists and section headers that can be shuffled or excerpted without breaking coherence
• Fewer unexplained metaphors, sarcasm, culture-specific humour
• Start with a short summary / TLDR
  • The first 2-3 lines are often harvested by RAG or vertical search
  • Also if someone uses an AI to summarise your piece, this helps ensure accuracy
     

1. ^{^}

   https://marginalrevolution.com/marginalrevolution/2025/01/should-you-be-writing-for-the-ais.html

2. ^{^}

   Search for "Generative Engine Optimization (GEO)" in this podcast