BenjaminTodd

Ug, we've been having a lot of problems with our new popup plugin. Once you're signed up, it should turn off all popups. Maybe you're blocking cookies, or it didn't take effect yet.

I agree it would be better to take you back to where you were reading, it's a lot more complex to set up though.

Sorry about that. I find popups annoying too. Unfortunately we just find we get a way higher conversion to the newsletter when we use popups, and since people need to engage over several months to get value out of the advice, getting people onto the newsletter is really important.

Hi everyone,

I'm the author of the piece. If you know any other good resources or statistics we should incorporate, I'm keen to hear about them. If you disagree with something, feedback is very welcome.

I'm the lead author of the piece and I agree there's a lot more work to do on these issues. That's why we set up 80,000 Hours.

We address one of your questions in a later article: how to work out what you're good at https://80000hours.org/career-guide/personal-fit/ (And beyond this, you need to make predictions - I think the book superforecasters is good on what to do there)

Which sign up requests in particular annoyed you?

Our focus is on careers with social impact. That's obvious if you come in from the main site, and we mention it on the first page of the tool, but it's easy to miss if you click straight through to the tool.

Hi all,

We've done one-on-one career coaching with a couple of hundred people, and found ourselves asking the same questions over and over again. So we put the most common questions into this simple tool.

It acts like a checklist to make sure you haven't missed a key bit of our advice. It's on the way to automated coaching.

It takes about an hour to work all the way through. That's long, but people use it for major life decisions.

We built it in Guidedtrack, a simple high-level language for making Q&A tools. It allows our researchers to edit the content quickly. We then overlaid our CSS theme. https://www.guidedtrack.com/

It gets about 1,000 users per week.

The aim of the tool is to cause people to change their career plans. 17% of users report a career plan change from engaging with us (though not necessarily from the tool itself).

4% make it to the end. That's less than 17% because we have a survey about a quarter of the way in, as well as at the end.

Net promoter score is -10%, and use isn't growing.

What are the most effective ways to make it better?

Good point. It still seems unlikely to me that all entrepreneurial activity contributes to productivity growth though. I'll look into it some more.

Quick aside - all the figures in the paper are per hour earnings figures.

Yes - I'd guess entrepreneurs have more social impact than salaried workers, because they do more innovative work. We'll write more about this in the future.

Note that the paper you cite is only about technological advances though, and only a fraction of what entrepreneurs do contributes to technological advances - a bunch of its just implementing stuff that's already best practice.

There's a lot of exploration of different development curves for AI in Bostrom's book, including the most obvious one, an S-curve.

Bostrom's predictions about when human level AI will arrive are mainly based on surveys of computer scientists e.g.

http://www.givewell.org/labs/causes/ai-risk/ai-timelines

I'm not saying we don't know whether climate change poses a tail risk or not (it obviously does). I'm just saying that claiming uncertainty isn't a good reason to avoid action.

In general, if there's a poorly understood but potentially very bad risk, then (a) more research to understand the risk is really high priority (b) if that research doesn't rule out the really bad scenario, we should try to do something to prevent it.

With AI, unfortunately waiting until the evidence that it's harmful is well established is not possible, because then it could be too late.

What AI risk research could involve is laid out in detail in the link.

The letter isn't (just) about modern weapon systems. It was put together by this group: http://futureoflife.org/ai-news/

Also, no-one is worried about a skynet scenario. The worrying scenario is just any powerful system that optimises for something that's different from what humans want.

Second, the point is that even uncertainty is enough for action. For AI to not be a problem, you'd need to be very confident that it'll occur a long way in the future, and that there's nothing we can do until it's closer. As you've said, we don't have confidence in the timeline. We have large uncertainty. And that's more reason for action, especially research.

Consider analogously:

"We've got no idea what the chance of run-away climate change is, so we shouldn't do anything about it."

Seems like a bad argument to me.

The expert consensus says 10% chance of human-level AI in 10 years: http://www.givewell.org/labs/causes/ai-risk/ai-timelines

Many computer science professors have publicly said they think AI poses significant risks. There's a list here: http://slatestarcodex.com/2015/05/22/ai-researchers-on-ai-ri...

Also see this open letter, signed by hundreds of experts: http://futureoflife.org/ai-open-letter/

The purpose of the profile isn't to argue a risk exists. We largely defer to the people we take to be experts on the issue, especially Nick Bostrom. We think he presents compelling arguments in Superintelligence, and although it's hard to say anything decisive in this area, if you think there's even modest uncertainty about whether AGI will be good or bad, it's worth doing more research into the risks.

If you haven't read Bostrom's book yet, I'd really recommend it. http://www.amazon.com/Superintelligence-Dangers-Strategies-N...

If you want to get a job in this area, we wrote a guide: https://80000hours.org/career-guide/top-careers/profiles/art...

In the spirit of openness, it would be great to see public responses to the downsides of this approach.

In particular, Bostrom Ch5.1 argues that the lead project is more likely than not to get a decisive strategic advantage, leading to a winner-takes-all scenario, which would mean attempts to foster a multipolar scenario (i.e. lots of similarly powerful AGIs rather than one) are unlikely to work.

In Ch11 he explores whether multipolar scenarios are likely to be good or bad, and presents many reasons to think they're going to be bad. So promoting the multipolar approach could be both very hard, and bad.

Thanks! So to summarise, you'll end up with lower earnings because your skills and network won't improve as fast as they would at a for-profit?

Point taken. The type of nonprofit I'm thinking about is the type that might go through YC though e.g. Watsi. How much would that change things?

Also, how about if you're product manager at the nonprofit (e.g. with a budget and doing some hiring) vs. just an engineer at a big firm?

To clarify the last point, you're saying the hiring market for junior engineers is especially great right now, so it might make sense to lock in the money while you can?

The list isn't in order.

The relevant difference is that animals feel pain but plants don't, because they don't have nervous systems.

We have to just pick one outcome variable for each experiment though, or it's going to get too complicated.

Sorry about that - we got a bit overloaded. It's also a beta platform, hence the loading page issue.

That's good feedback, thank you.

That would be better, though quite a bit more effort!

Usually these studies only have enough resolution to say "probably worked", "no measurable effect", "probably negative", and even then are often wrong.

Some have, but many persist for years (e.g. Scared Straight).

It's not evident from this particular post, but overall we put a lot of emphasis on the "career capital" you get out of different options (the extent to which they put you in a better position to have a positive impact in the future).

https://80000hours.org/career-guide/basics/

We also think "corruption risk" should be taken very carefully, which is why we advocate making a public pledge of your intention to donate, and being part of a community.

We also don't think you should take a high-earning career if it causes significant harm.

Try reading the career guide: https://80000hours.org/career-guide/basics/

I agree that's an issue - we're working on better mobile optimisation.

Our framework and individual career research, like it says at the top of the page.

https://80000hours.org/career-guide/basics/ https://80000hours.org/career-guide/profiles/