Confidence: Medium, underlying data is patchy and relies on a good amount of guesswork, data work involved a fair amount of vibecoding.  Intro:  Tom Davidson has an excellent post explaining the compute bottleneck objection to the software-only intelligence explosion.[1] The rough idea is that AI research requires two inputs: cognitive labor and research compute. If these two inputs are gross complements, then even if there is recursive self-improvement in the amount of cognitive labor directed towards AI research, this process will fizzle as you get bottlenecked by the amount of research compute.  The compute bottleneck objection to the software-only intelligence explosion crucially relies on compute and cognitive labor being gross complements; however, this fact is not at all obvious. You might think compute and cognitive labor are gross substitutes because more labor can substitute for a higher quantity of experiments via more careful experimental design or selection of experiments. Or you might indeed think they are gross complements because eventually, ideas need to be tested out in compute-intensive, experimental verification.  Ideally, we could use empirical evidence to get some clarity on whether compute and cognitive labor are gross complements; however, the existing empirical evidence is weak. The main empirical estimate that is discussed in Tom's article is Oberfield and Raval (2014), which estimates the elasticity of substitution (the standard measure of whether goods are complements or substitutes) between capital and labor in manufacturing plants. It is not clear how well we can extrapolate from manufacturing to AI research.  In this article, we will try to remedy this by estimating the elasticity of substitution between research compute and cognitive labor in frontier AI firms.  Model  Baseline CES in Compute To understand how we estimate the elasticity of substitution, it will be useful to set up a theoretical model of researching better alg

TLDR: This 6 million dollar Technical Support Unit grant doesn’t seem to fit GiveWell’s ethos and mission, and I don’t think the grant has high expected value. Disclaimer: Despite my concerns I still think this grant is likely better than 80% of Global Health grants out there. GiveWell are my favourite donor, and given how much thought, research, and passion goes into every grant they give, I’m quite likely to be wrong here!   What makes GiveWell Special? I love to tell people what makes GiveWell special. I giddily share how they rigorously select the most cost-effective charities with the best evidence-base. GiveWell charities almost certainly save lives at low cost – you can bank on it. There’s almost no other org in the world where you can be pretty sure every few thousand dollars donated be savin’ dem lives. So GiveWell Gives you certainty – at least as much as possible. However this grant supports a high-risk intervention with a poor evidence base. There are decent arguments for moonshot grants which try and shift the needle high up in a health system, but this “meta-level”, “weak evidence”, “hits-based” approach feels more Open-Phil than GiveWell[1]. If a friend asks me to justify the last 10 grants GiveWell made based on their mission and process, I’ll grin and gladly explain. I couldn’t explain this one. Although I prefer GiveWell’s “nearly sure” approach[2], it could be healthy to have two organisations with different roles in the EA global Health ecosystem. GiveWell backing sure things, and OpenPhil making bets.   GiveWell vs. OpenPhil Funding Approach What is the grant? The grant is a joint venture with OpenPhil[3] which gives 6 million dollars to two generalist “BINGOs”[4] (CHAI and PATH), to provide technical support to low-income African countries. This might help them shift their health budgets from less effective causes to more effective causes, and find efficient ways to cut costs without losing impact in these leaner times. Teams of 3-5

Crossposted on Substack. Preamble I have been asked to share a few tips on how to get the most out of the conference as part of the opening talk of this year’s EAGxPrague. I decided to focus on the tacit knowledge I have. This is because there is a lot of useful advice on how to get the most out of conferences, but I feel they usually offer general tips, as opposed to really getting into the weeds. Here is a slightly more polished version of the transcript I prepared. I also added some pictures: Hi, my name is Gergő, and I would like to welcome you to this conference too. Props to Hana for organising! Yesterday, I decided to check Swapcard to see how many EA conferences I attended since my first one in 2020. It’s around 20, so I thought I would exactly describe what I do and hope that I don’t come across as insane to you. Subscribe to The Field Building Blog Conference week Monday/Tuesday At the beginning of conference week, I first scroll through the agenda and add a couple of workshops and talks that I might be interested in. There are usually lot’s of interesting sessions, but I try to be selective as I know 1-1s are usually more valuable. At this point, I roughly end up with something like 6 sessions. This doesn’t take long. The reason it’s good to do it early because most people haven’t yet started networking. For the times I booked sessions, I set myself unavailable so people can’t book me at those slots. I wish Swapcard did this automatically, but it doesn't.         Wednesday/Thursday I set aside time on at least part of my Wednesday or Thursday to continue preparing. I do this on one of these days because, by then, most attendees have registered for the conference. Before this time, their profiles won’t show up on Swapcard. This is where the heavy lifting starts. I spend at least 3, but depending on the size of the conference, more than 6 hours networking on Swapcard. Figuring out who to talk to I literally scroll through every profile on