Alexander Berger

How to Sound Like an Expert in Any AI Bubble Debate There are 12 statistics, factoids, and studies that dominate every discussion about whether artificial intelligence is a bubble. Here's a deep-dive into all 12 arguments

Every AI-bubble debate is the same 12 facts, studies, and arguments quoted back and forth, over and over.

So Tim Lee at @understandingai.skystack.xyz and I teamed up to publish a 4,000 word deep-dive analysis of all 12.

www.derekthompson.org/p/how-to-sou...

Here's the link to the full essay again: ssir.org/articles/en...

... but also preserves some space for pluralism.

These types of judgment calls are among the most uncertain ones we make; we try to strike a balance between maximization and maintaining a commitment to pluralism across worldviews. I'm sure that balance will continue to evolve as we learn more.

For allocating across very different kinds of value, instead of trying to force everything into the same terms and then explicitly maximize, we tend to allocate budget based on how holistically attractive we find the relevant worldview - that incorporates some sense of how cost-effective it looks...

Within any given portfolio area, we try to optimize as much as possible. But trying to maximize across portfolios can easily go way too far. My co-founder Holden wrote about the perils of maximization in a 2022 post that I still think about often:
forum.effectivealtruism.org/posts/T975y...

Most potential answers lead to all or nothing allocations:

- If you think averting a year of animal suffering is similar in value to a year of human life, you’ll only work to avert animal suffering.
- If you think no amount of animal suffering compares to a year of human life, you’ll do none.

In a world of scarcity, the logic of maximizing cost-effectiveness pushes to have some shared unit of comparison, where we can say how to split our budget between helping animals and humans, or how many lives today are worth some degree of risk to the future of humanity overall.

The causes we work in each represent an extraordinary opportunity to do good, and each could be all-consuming:

- In global health, some interventions can save a life for ~$5k
- A single dollar can avert 10 years of animal suffering
- Some grantees are literally trying to prevent human extinction.

The fourth lesson is really more of an ongoing tension we navigate, between trying to maximize expected value to help others as much as we can and maintaining a commitment to moderation and pluralism.

The third lesson: when we started, I thought the best giving opportunities would be “safe” bets like GiveWell recs or low-probability, high-upside bets to reduce catastrophic risks. I now think you can get just as strong returns with medium-risk opportunities, like R&D and advocacy:

Alexander Berger (@albrgr.bsky.social) The strategic landscape was very unclear when we first entered the field. As a result, we mostly funded early-stage research and field-building efforts to increase the number of people taking these questions seriously.

We've also seen this with AI safety and security: when we began funding in the space, the field was solidly pre-paradigmatic. It has grown enormously since then and is much better able to meet demand from companies and policymakers
bsky.app/profile/alb...

Alexander Berger (@albrgr.bsky.social) CA Governor Newsom just signed SB 79, an important bill to make homes easier to build near transit, and arguably the biggest YIMBY win to date. Some reflections on the bill & the history of how we got here 🧵

The second lesson: there are huge returns to investing in early field-building, though it can take time to pay off.

I've seen this with the recent successes of the YIMBY movement, which we began supporting in 2015:
bsky.app/profile/alb...

Strategic Cause Selection | Coefficient Giving Philanthropists tend to choose causes based on personal experience, geographic proximity, or emotional appeal. This approach is understandable — personal connection is what motivates many people to give in the first place. But some causes offer far greater opportunities for impact than others: the same donation might save ten lives in one place but a thousand in another.

The first lesson is something we talk a lot about: everything you do is downstream of the causes you select, so considering a wide range of causes and thinking rigorously about where you can have the most impact is crucial.

Here's our longer take: coefficientgiving.org/research/st...

Today, I have an article in @SSIR.org on lessons from our first decade, including:

(1) The importance of choosing causes
(2) Field-building pays off hugely
(3) Great philanthropy is possible anywhere on the risk spectrum
(4) Balancing maximization and moderation is hard

🧵
ssir.org/articles/en...

Great to see the US renewed funding for the Global Fund! www.devex.com/news/global...

Investment in tech equip & software rivals the 90s in growth rates but accounted for 2/3 of all growth in H1 vs only ~15% in the late 90s. Other things going on in the 90s--broad based boom in the labor market/GDP, globalization, peace dividend, consumer optimism, labor force growth & immigration

More on our big news today:
bsky.app/profile/alb...

I’m extremely grateful to my colleagues and to the donors and grantees who have made this new chapter possible. They've set us up to deliver far more impact in our second decade than in our first. I can't wait to see what we accomplish together.

We’re incredibly excited for what more effective philanthropic capital could unlock: new vaccines for some of the world’s deadliest diseases, intelligent navigation of rapid AI progress, faster economic growth, and meaningful reductions in animal suffering.

Partner With Us | Coefficient Giving

Our work with donors also includes creating custom menus of giving opportunities, “learning journeys” that bring donors up to speed on areas of interest, and operational advising and support. Learn more here:
coefficientgiving.org/about-us/pa...

Funds | Coefficient Giving Our funds enable philanthropists to pool resources around rigorously selected, high-impact cause areas.

From today, all our programs will work similarly to LEAF and AGF - they’ll be funds that other donors can support at scale.

Each will seek to pool capital and direct it toward the most promising opportunities in a cause area. Full list here, now with pretty colors:
coefficientgiving.org/funds/

Alexander Berger (@albrgr.bsky.social) Exciting update: @open_phil is doubling down on our YIMBY, innovation, and metascience success by launching a >$120m Abundance & Growth Fund to accelerate economic growth and boost scientific & technological progress. Funding from @goodventures.bsky.social, Patrick Collison, and others. Why now? 🧵

Another example: this year, we launched the >$120m Abundance & Growth Fund, led by @mattsclancy.bsky.social, to boost economic growth and scientific & technological progress.
bsky.app/profile/alb...

Lead Exposure Action Fund | Coefficient Giving

After making several promising exploratory grants alongside GiveWell, another donor we've worked with before suggested we start a pooled fund to scale this work.

Today, LEAF supporters have collectively committed over $125m to reduce lead poisoning.
coefficientgiving.org/funds/lead-...

Reducing lead poisoning scored well on three criteria we use as a proxy for cost-effectiveness:

- Importance: 1m deaths per year, >⅓ of children worldwide affected
- Neglectedness: <1% the funding of malaria or TB
- Tractability: Proven ways of reducing lead in spices & paint

Alexander Berger (@albrgr.bsky.social) Great new article on the tragedy of lead exposure: - 1 in 3 kids have high levels of lead in their blood - Some estimates attribute 5.5m deaths (more than AIDS+TB+malaria) - Costs ~1% in global GDP This is an important, tractable, and neglected problem. We can make progress! https://bsky.app/profile/leecrawfurd.bsky.social/post/3lrfvqxr5hj23

An illustrative example of our work with other donors is our Lead Exposure Action Fund, which has brought together 12 philanthropists to more than double baseline global spending on lead reduction in poor countries.
bsky.app/profile/alb...

The Story Behind Our New Name | Coefficient Giving A coefficient multiplies the value of whatever it’s paired with, just as we aim to amplify impact through our research, grantmaking, and partnerships. We knew that any name would take some getting used to, but this one captured something essential about our work.

Our new name reflects various aspects of this new chapter:

“Co” -> collaborating with other donors
“Efficient” -> a nod to cost-effectiveness
“Coefficient” -> multiplying others’ impact, ROI

(And “giving” is much less of a mouthful than “philanthropy”)
coefficientgiving.org/research/th...

Our ambition has always been to work with more donors once we had enough bandwidth to support Good Ventures.

We started in earnest in 2024, directing over $100m from other donors. We more than doubled that so far in 2025. We’re aiming for a lot more in years to come.

In the coming years, we’ll continue to grow our work with Good Ventures. But many others are interested in giving at scale and not on track. We often hear people cite a lack of outstanding, shovel-ready opportunities.

Over the last decade, we’ve worked closely with our founding partner Good Ventures to give away over $4b. I am very proud of the results:

Open Philanthropy Is Now Coefficient Giving | Coefficient Giving Our new name marks our next chapter as we double down on our longstanding goal of helping more funders increase their impact. We believe philanthropy can be a far more vital force for progress than it is today.

First, here’s a link to my essay about the shift (on our brand new website, click around!) and what we're trying to do in our second decade:
coefficientgiving.org/research/op...