Our strategy at 80,000 Hours

What we do

80,000 Hours provides research, information, and support to help talented people move into careers that tackle the world’s most pressing problems.

Our strategic focus

What we provide

Our programmes help us to achieve our strategic focus. Each programme has a team dedicated to it. Here’s a brief summary of what they do:

Support teams

In addition to our programmes, which create resources for our audience, we also have the following cross-functional supporting teams:

Why we exist

Our broader organisational identity

Our problem, vision, and mission — underlying our focus on making advanced AI go well — is:

PROBLEM: Many talented people want to do good with their careers, but existing advice rarely covers how to do so effectively, meaning they have far less impact than they could.
VISION: Get more talented people effectively tackling the world’s most pressing problems.
MISSION: Provide free research, information, and support to help people find high-impact careers.

80,000 Hours is cause-neutral: our cause prioritisation, strategy, and activities would change if we no longer thought that making advanced AI go well was the most pressing global problem.

Why impact-oriented careers advice?

1. Lots of people want to do good.
2. People spend a huge fraction of their time on their career, so it’s a strong leverage point for having an impact (even in worlds where AI research and development is automated in the next few years).
3. There are huge differences in impact between careers, but existing advice rarely makes comparisons of impact.
4. We think that providing career advice that makes comparisons of impact can reliably lead to more people tackling the biggest problems more effectively. It’s also neglected and scalable.
5. So, we think it’s one of the best ways to make progress on pressing global problems.
6. And we think that we have an empirical track record of helping people move into paths we think are highly impactful.

How we think about cause prioritisation

Here are some of the foundational assumptions we are making when thinking about how we can have the biggest impact:

• We care about total wellbeing, considered appropriately impartially — that is, welfare of beings, regardless of location or substrate,¹ over the long term. (Read more about how we think about social impact.)
• We think considering the effects of an action or state of affairs in expectation is an important and underrated way to deal with uncertainty, even though we do not take expected value calculations literally. (Read more about expected value.)
• Improving the prospects for all future generations is among the most morally important things we can do. (Read more about longtermism.)
• We avoid violating common-sense moral constraints or naively maximising for promoting a particular conception of moral value. (Learn more about the perils of naive maximisation)
• Given moral and empirical uncertainty, we aim to consider a range of perspectives and possibilities, and take the actions that seem best on balance. (Read more about moral uncertainty.)
• A key framework we use when assessing the pressingness of different global problems is to assess their importance, tractability, and neglectedness. (Read more about our decision framework.)

Why we are focused on making advanced AI go well

In addition to the foundational assumptions above, we have empirical views which cause us to focus in particular on risks (and opportunities) stemming from advanced AI:

1. AI progress is continuing rapidly. We think that likely by 2040,² and potentially much sooner, AI systems will exist that can do the most economically valuable cognitive tasks (scientific research, designing new technologies and products, running businesses, etc).
2. Highly capable AI systems could rapidly transform the world and pose challenges we aren’t prepared for. For example, they could make it easier than ever for humans to launch successful coups, and to design extremely powerful weapons. Or they could have goals of their own, and seek to disempower humanity in order to achieve them.
3. The stakes are potentially existential. Advanced AI could bring unprecedented prosperity, but it could also lead to humanity going extinct or losing the ability to shape its own future.
4. The issue is neglected relative to its importance: only a few thousand people work on it compared to the millions working on conventional social issues.
5. It’s tractable now: laws, norms, and institutions around AI are still being formed, which means people have an opportunity to shape them.

How we work

Cultural values

At present, our top cultural values are:

1. Ambitious long-term impact
2. A modest, scientific mindset
3. Openness and honesty
4. Fostering a fast, innovative, AI-focused internal culture.

At a high level, how do we achieve our strategic focus and mission?

1. Develop views on which interventions and roles matter most for making advanced AI go well.
2. Reach people through engaging content, marketing, and word-of-mouth.
3. Communicate why and how to make advanced AI go well — introducing people to the most important ideas and helping them see how they can contribute, especially via their careers.
4. Get people into roles that help make advanced AI go well through career advice, headhunting, and our job board.
5. Help people stay up to date so they can adjust their plans as the world changes.

How programme strategy is set at 80k

80,000 Hours has an organisation-wide strategy, and individual programme strategies flow from and contribute to this. Programme strategies are set at the programme level, but the CEO works with each programme to select a strategy and plan that contributes to the organisational strategy. The CEO is ultimately responsible for deciding whether to start, spin out, or shut down programmes.

Centralised structures

In addition to the cross-functional supporting programmes, we have some centralised structures to support the programmes. For example:

• Programmes collaborate much more closely than separate organisations would, such as through:
  • Fortnightly senior leadership team meetings to coordinate on priorities, collaboration, and organisation-wide questions
  • Cross-team support, shared policies, a shared brand, and feedback
    • Some concrete examples: anyone can give feedback on advising calls or podcast episodes; programmes cross-promote each other; hiring committees must contain one cross-team member; etc.
• Impact evaluation is done centrally (and sometimes also by individual programmes). We use it to provide a feedback loop that allows us to answer strategic questions, improve the quality of our products, and track whether we’re succeeding in our mission.
  • Read a summary of our historical impact here.

Target audience

Programmes set their own specific target audiences, but broadly we target:

• English-speaking students and professionals around the world, typically aged 18–45, who are unusually high in ability and openness to engaging with questions with a scope-sensitive, impartially-altruistic, collaborative, and truth-seeking mindset.

Currently, the website mostly targets the younger end of this age range, while our career services, video, and studio teams try to span the whole range. We also think of LLMs that serve this audience as in our target audience.

Strategic assumptions informing our choice of target audience

• The distribution of people’s career impact is often heavy-tailed in the most important areas.
• People who end up having a lot of impact over the course of their career are more likely to have a lot of potential when young, and tend to be open to engaging with questions in a way that is scope-sensitive, impartially-altruistic, collaborative, and uses a scout mindset.
• We can attract people who are very high on these criteria and engage them.
• Younger people tend to be more willing to change careers and do unusual things.
• Older people can often contribute to our pressing problems sooner than younger people.
• When deciding between engaging with individual users (e.g. on career services), we prioritise based on how much impact we think we can have by helping them have a higher impact career.³

Relationship with Effective Altruism (EA)

We consider ourselves part of the EA ecosystem, but we aim to act on our own views and appeal to our much wider target audience. Our current audience is also much wider than the EA community. (Read more about our relationship with effective altruism.)