Co-founder and CEO of Probably Good.
Also a co-founder of EA Israel.
Co-founder and CEO of Probably Good.
Also a co-founder of EA Israel.
Thank you Michelle for posting this!
I wanted to add that when we founded Probably Good, we were also worried of exactly the sort of things detailed here. The reality turned out to be much better than we had even hoped: We contacted 80k and everyone on the team were incredibly supportive. This is both true in their attitude (being happy to see more people in this space) and also in practice (with advice and help along the way).
So I’m very happy to say that I agree. There’s still lots of areas to explore in EA career advice and the people at 80k are far far nicer than you’d expect from someone when you tell them you want to ‘compete’* with them.
Besides that, naturally, if we (Probably Good) can help people with advice or anything else, we’d be more than happy to!
* I know, it’s not really competing. We really are genuienly working towards common goals.
Thank you for writing this!
I think your analysis can be specifically useful for people who want to contribute and feel like they’re not sure where to look for neglected areas in EA.
I’ll add a small comment regarding “It is difficult to compete with the existing organisations that are just not quite doing this”:
My experience with orgs in the EA community is that pretty much everyone is incredibly cooperative and genuinely happy to see others fill in the gaps that they’re leaving.
I’ve been in talks with 80,000 hours and a few other orgs about an initiative in the careers space for a while now. Everyone we’ve talked to was both open about what they’re doing (and what they aren’t doing) and ridiculously helpful with advice and support.
I think if someone is serious about trying to fill a gap in the EA body of work: It’s important to understand from adjacent orgs how big \ real this gap is and if they have comments about your approach to it. And while I can see why someone would be worried, I think if you approach with the right attitude, the ‘competition’ would have far more benefits than harms.
I generally agree with the idea and appreciate the clarity of this post.
One related thought which I think is potentially useful both for thinking of which projects to fund or to start:
Projects usually need to be scalable at advanced stages, but not at the start. It’s ok (and even recommended* in many cases) to start doing things in non-scalable ways that aren’t cost-effective.
A lot of times, the value in information \ experience \ growth is high enough that it’s worth starting out doing things that you won’t be able to sustain as you grow.
Obviously, there should be a plan (or at least ideas on how) to become more scalable later. I’d be looking for projects that have reasonable path(s) to being very scalable down the road.
* This link is advice for for-profit startups. It’s only partially relevant for our context but the point I’m making is made there in more detail.
First of all, thank you for the feedback! It’s not always easy to solicit quality (and very thoroughly justified) feedback, so I really appreciate it.
Before diving into the specifics, I’ll say that on the one hand—the name could definitely change if we keep getting feedback that it’s suboptimal. That could be in a week or in a year or two, so the name isn’t final in that sense.
On the other hand, we did run this name quite a few people (including some who aren’t familiar with EA). We tried (to the best of our ability) to receive honest feedback (like not telling people that this is something we’re setting up or letting someone else solicit the feedback). Most of what you wrote came up, but rarely. And people seemed to feel positively about it. It’s definitely possible that the feedback we got on it was still skewed positive, but it was much better for this name than for other options we tried.
Now, to dive into the specifics and my thoughts on them:
* The name doesn’t make the function clear: I think this is a stylistic preference. I prefer having a name that’s more memorable, when the function can be explained in a sentence or two right after it. I know the current norm for EA is to name orgs by stating their function in 2 or 3 words, but I think the vast majority of orgs (for profit and non-profit) choose a name that doesn’t just plainly state what the org does. I will mention that, depending on context, what might appear is “Probably Good Career Advice”, which is clearer (though still doesn’t fully optimize for clarity).
* Good can mean quality and morality: Again, I liked that. We do mean it in both ways (the advice is both attempting to be as high quality as possibly and as high as possible in moral impact, but we are working under uncertainty in both parameters).
* Turning people off by giving the message that the product isn’t good or that we’re not ambitious in making it good: I pretty much fully agree with you on the analysis. I think this name reduces the risk of people expecting a level of certainty that we’ll never reach (and is very commonly marketed in non-EA career advice) and increases the risk of people initially being turned off by perceived low quality or low effort.
I also like and agree with your “pitch” and that is more of less how I’m thinking about the issue.
Two relevant points on weighing this trade-off:
1. Currently, I’m more worried about setting too high expectations than the perception of low quality. Both because I think we can potentially cause more harm (people following advice with less thought than needed) and because I think there are other ways to signal high quality and very few ways to (effectively) lower people’s perceived certainty in our advice.
2. Most people we ran the name by did catch on that the name was a little tongue-in-cheek in it’s phrasing. This wasn’t everyone, but the people who did see that—didn’t think there was a signal of lower quality.
I do agree there’s a risk there, but I see it as relatively small, especially if I’m assuming that most people will reach us through channels where they have supposedly heard something about us and aren’t only aware of the name.
To summarize my thoughts:
I don’t think it’s a perfect name.
I like that it’s a memorable phrase rather than a bland statement of what we do. I like that it’s a little tongue-in-cheek and that it does a few things at the same time (two meaning or good, alluding the the uncertainty). I like that it put our uncertainty front and center.
I agree there’s a risk of signaling low quality \ effort and that all of the things that I like could also be a net harm if I’m wrong (which isn’t specifically unlikely).
We’ll collect more feedback on the name and we’ll change if it doesn’t look good.
Thanks for asking!
There are a few ways we distinguish ourselves, and hopefully complement 80K’s work.
Though the cause areas we care about are not mutually exclusive, we often focus on different cause areas (e.g. we’ve done a lot of work in global health and development, and have a talk about this work at EAG London—you’re welcome to come to the talk to hear more).
We aim to cater to a broader audience in terms of worldviews, geographic diversity, and educational background.
There are many differences in how we approach career advice, the most significant is that we are a lot more uncertain about things (as indicated by the name Probably Good) - this influences how we make recommendations, how we engage with advisees, what career paths we’re open to, and more.
Finally, and most importantly, it’s worth noting that we believe that impact-focused career advice is an incredibly large and important field, so we think it’s good to have multiple teams taking different approaches to it. Hence our primary drive for our work at Probably Good is not to be different from 80K (who are great and have been both an inspiration and a huge help in the creation of Probably Good), but rather to do more without necessarily trying to be similar.
Thanks for writing this! It’s always useful to get reminders for the sort of mistakes we can fail to notice even if when they’re significant.
I also think it would be a lot more helpful to walk through how this mistake could happen in some real scenarios in the context of EA (even though these scenarios would naturally be less clear-cut and more complex).
Lastly, it might be worth noting the many other tools we have to represent random variables. Some options off the top of my head:
* Expectation & variance: Sometimes useful for normal distributions and other intuitive distributions (eg QALY per $ for many interventions at scale).
* Confidence intervals: Useful for many cases where the result is likely to be in a specific range (eg effect size for a specific treatment).
* Probabilities for specific outcomes or events: Sometimes useful for distributions with important anomalies (eg impact of a new organization), or when looking for specific combinations of multiple distributions (eg the probability that AGI is coming soon and also that current alignment research is useful).
* Full model of the distribution: Sometimes useful for simple \ common distributions (all the examples that come to mind aren’t in the context of EA, oh well).
One small note: The examples are there to make the category clearer. These aren’t all cases where expected value is wrong \ inappropriate to use. Specifically, for some of them, I think using expected value works great.
By the way, both Sella and I (Omer) will be at EAG. Feel free to connect, we’d love to chat with people who are interested in Probably Good and hear what you have to say...
For the sake of clarity I’ll restate what I think you meant:
We’re not discussing the risk of people taking less impactful career paths than they would have taken counterfactually because we existed (and otherwise they might have only known 80k for example). That is a risk we discuss in the document.
We’re talking specifically about “membership” in the EA community. That people who are less committed \ value aligned \ thoughtful in the way that EAs tend to be \ something else—would now join the community and dilute or erode the things we think are special (and really good) about our community.
Assuming this is what you meant, I’ll write my general thoughts on it:
1. The extent to which this is a risk is very dependent on the strength of the two appearances “very” in your sentence “A career org that (1) was very broad in its focus, and/or very accepting of different views”. While we’re still working out what the borders of our acceptance are (as I think Sella commented in response to your question on agnosticism), we’re not broadening our values or expectations to areas that are well outside the EA community. I don’t currently see a situation where we give advice or a recommendation that isn’t in line with the community in general. It’s worth noting that the scope and level of generality that the EA community engages with in most other interactions (EA Global, charity evaluation orgs, incubation programs, etc.) is much broader than 80K’s current focus. We see our work as matching that broader scope rather than expanding it, and so we don’t believe we’re changing where EA stands on this spectrum—simply applying it to the career space as well.
2. More importantly, even in cases where we could make a recommendation that (for examples) 80k wouldn’t stand behind—our methodology, values, rigor in analysis, etc. should definitely be in line with what currently exists, and is expected, in the community. I can’t promise we won’t reach different conclusions sometimes, but I won’t be “accepting” of people who reach those conclusions in shoddy ways.
3. This is a relatively general point, but it’s important and it mitigates a lot of our risks: In the next few months, we’re not planning to grow, do extensive reaching out, market or try to bring a lot of new people in. That’s explicitly because we want to create content and start working, do our best to evaluate the risks (with the help of the community) - and only start having a large impact once we’re more confident in the strength and direction of that impact.
In a sense (unless we fail pretty badly at evaluating in a few months) - we’re risking a very small harm of a small unknown org and potentially gaining the benefits that could be quite large if we do find that our impact looks good.
Thank you for writing this!
I really appreciate your approach of thoroughly going through potential issues with your eventual conclusion. It’s a really good way of getting to the interesting parts of the discussion!
The area where I’m left least convinced by is the use of Laplace’s Law of Succession (LLoC) to suggest that AGI is coming soonish (that isn’t to say there aren’t convincing arguments for this, but I think this argument probably isn’t one of them).
There are two ways of thinking that make me skeptical of using LLoC in this context (they’re related but I think it’s helpful to separate them):
1. Given a small amount of observations, there’s not enough information to “get away” from our priors. So whatever prior we load into the formula—we’re bound to get something relatively close to it. This works if we have a good reason to use a uniform prior or in contexts where we’re only trying to separate hypotheses that aren’t “far enough away” from the uniform prior, which I don’t think is the case here:
In my understanding, what we’re really trying to do is separate two hypotheses: The first is that the chance of AGI appearing in the next 50 years is non-negligible (it won’t make a huge difference to our eventual decision making if it’s 40% or 30% or 20%). The second is that it is negligible (let’s say, less than 0.1%, or one in a thousand).
When we use a uniform prior (which starts out with a 50% chance of AGI appearing within a year) - we have already loaded the formula with the answer and the method isn’t helpful to us.
2. In continuation to the “demon objection” within the text, I think the objection there could be strengthened to become a lot more convincing. The objection is that LLoC doesn’t take the specific event it’s trying to predict into account, which is strange and sounds problematic. The example given turns out ok: We’ve been trying to summon demons for thousands of years so the chance of it happening in the next 50 years is calculated to be small.
But of course, that’s just not the best example to show that LLoC is problematic in these areas:
Example 1: I have thought up of a completely new and original demon. It was obviously never attempted to summon my new and special demon until this year, when, apparently it wasn’t summoned. The LLoC chance of summoning my demon next year is quite high (and over the next 50 years is incredibly high). It’s also larger than the chance of summoning any demon (including my own) over those time periods.
The problematic nature of it isn’t just because I picked an extreme example with a single observation -
Example 2: What is the chance that the movie Psycho is meant to hypnotize everyone watching it and we’ll only realize it when Hitchcock takes over the world? Well, turns out that this hasn’t yet happened for exactly 60 years. So, it seems like the chance of this happening soon is precisely the same as the chance of AGI appearing.
Next, what is the chance of Hitchcock doing this AND Harper Lee (To Kill a Mockingbird came out in the same year) attempts doing this in a similar fashion AND Andre Cassagnes (Etch-A-Sketch is also from 1960) does so (I want to know the chance of all three happening at the exact same time)? Turns out that this specific and convoluted scenario is just as likely since it could only start happening at 1960… This is both obviously wrong and an instance of the conjunction fallacy.
My initial intuition (stressing even more that this is based on no evidence but my best guess) is that the name “Probably Better” would be more confusing to people than “Probably Good”. I’m expecting a lot of people asking “better than what?”
It also loses the meaning of good as in moral good (which I like, but not everyone here did).
Thank you for writing what you’d find most valuable! This lines up well with my thoughts...
Regarding being overwhelmed by requests for advice:
Yes! That’s definitely a failure mode. We’ve discussed how much we can give direct advice (very little in the near future, potentially more later but that’s quite a bit of work to get there) and how to choose candidates (where we have a lot of thoughts but, as with other things, we expect to decide on a criteria and then have to fix it once we see where it fails).
I’m cautiously optimistic that we just don’t have enough time to fall into this failure mode and so we’ll stop ourselves before this becomes an issue :-)
Wow! This is really good!
I think the general advice is great, and I really appreciate your candidness: Revealing the data and the materials you used, as well as the level of detail regarding your process.
This isn’t something that is usually written and I’m sure it’ll help a lot of people facing hiring challenges for EA orgs...
I can add a little about my own experience and process regarding rejection (which I agree is one of the hardest parts):
1. I try to honestly explain to candidates why they were rejected (usually by mail, sometimes by phone). This is usually possible for almost any candidate who has had an interview that wasn’t very short (with the exception of a few candidates that I have a very strong impression that they don’t want to hear it). Specifically, if possible, I try to answer the question of “What would need to change for you to be accepted in a year”. I started out very nervous about how candidates will receive it and have been surprised at how much it’s appreciated.
2. I really agree with what you wrote about not being a jerk, and that timely answers are an important part of it. This is especially true for rejection, partly because it’s easy for us to procrastinate making a decision when that decision is uncomfortable.
3. I think it’s important to make everything is worded precisely and clearly and leaves no room for misinterpretation. Be careful not to give false hope that you might still reconsider (if that’s not true), don’t write something that might be interpreted as hinting at some hidden reasons for the rejection, etc. This isn’t the place for writing with style. It should be optimized for conciseness and clarity. This is also why I usually send rejections by email, rather than phone. Phone calls are more personal but I can look over what I write in an email and make sure it says exactly what I mean.
4. In cases where I have a good impression of a candidate but there isn’t a fit, I offer to intro them to people I know who are also hiring for similar roles at other orgs\companies. It’s a good way of helping everyone involved and shows that I really do believe they can be great for other roles\orgs.
Yes, that makes perfect sense. I think we definitely need to have a system that (1) let’s people know if they’re not going to get coaching even though they asked and (2) doesn’t take up a lot of our time.
Thank you for the input!
I think some of the questions you raised are (at least partially) answered in our documents. Specifically, where we detail the impacts that we hope to achieve—those are impacts that we think we would potentially have a comparative advantage over 80,000 hours. Areas where we think we would be similar to 80,000 hours wouldn’t be areas where we’d expect to have significant counterfactual impact.
Regarding the abstractness and general nature of the documents, that’s completely fair. I expect things will be a lot clearer when we have a website up and some content, rather than documents explaining the principles by which we are creating the content.
As we’ve written in a few places, we’re taking this one step at a time and trying to get as much feedback as possible at every stage. I hope it won’t be very long before we’re able to start publishing some of our materials, which will be a good example of our actual work and will convey the specifics of our focus.
That sounds really cool!
I’ll be happy to join! :-)
I wanted to chime in and say that while a lot of people / organizations say things like this, in my experience, 80,000 really does mean it and follows through. When we were setting up Probably Good (and not only then) the amount of encouragement and help we received from Michelle, Niel and others there has been incredible.