Maybe I’m being cynical, but I’d give >30% that funders have declined to fund AI Safety Camp in its current form for some good reason. Has anyone written the case against? I know that AISC used to be good by talking to various colleagues, but I have no particular reason to believe in its current quality.
MATS has steadily increased in quality over the past two years, and is now more prestigious than AISC. We also have Astra, and people who go directly to residencies at OpenAI, Anthropic, etc. One should expect that AISC doesn’t attract the best talent.
If so, AISC might not make efficient use of mentor / PI time, which is a key goal of MATS and one of the reasons it’s been successful.
Why does the founder, Remmelt Ellen, keep linkposting writing by Forrest Landry which I’m 90% sure is obvious crankery? It’s not just my opinion; Paul Christiano said “the entire scientific community would probably consider this writing to be crankery”, one post was so obviously flawed it gets −46 karma, and generally the community response has been extremely negative. Some AISC work is directly about the content in question. This seems like a concern especially given the philosophical/conceptual focus of AISC projects, and the historical difficulty in choosing useful AI alignment directions without empirical grounding. [Edit: To clarify, this is not meant to be a character attack. I am concerned that Remmelt does not have the skill of distinguishing crankery from good research, even if he has substantially contributed to AISC’s success in the past.]
All but 2 of the papers listed on Manifund as coming from AISC projects are from 2021 or earlier. Because I’m interested in the current quality in the presence of competing programs, I looked at the two from 2022 or later: this in a second-tier journal and this in a NeurIPS workshop, with no top conference papers. I count 52 participants in the last AISC so this seems like a pretty poor rate, especially given that 2022 and 2023 cohorts (#7 and #8) could both have published by now. (though see this reply from Linda on why most of AISC’s impact is from upskilling)
The impact assessment was commissioned by AISC, not independent. They also use the number of AI alignment researchers created as an important metric. But impact is heavy-tailed, so the better metric is value of total research produced. Because there seems to be little direct research, to estimate the impact we should count the research that AISC alums from the last two years go on to produce. Unfortunately I don’t have time to do this.
MATS has steadily increased in quality over the past two years, and is now more prestigious than AISC. We also have Astra, and people who go directly to residencies at OpenAI, Anthropic, etc. One should expect that AISC doesn’t attract the best talent.
If so, AISC might not make efficient use of mentor / PI time, which is a key goal of MATS and one of the reasons it’s been successful.
AISC isn’t trying to do what MATS does. Anecdotal, but for me, MATS could not have replaced AISC (spring 2022 iteration). It’s also, as I understand it, trying to have a structure that works without established mentors, since that’s one of the large bottlenecks constraining the training pipeline.
Also, did most of the past camps ever have lots of established mentors? I thought it was just the one in 2022 that had a lot? So whatever factors made all the past AISCs work and have participants sing their praises could just still be there.
He was posting cranky technical stuff during my camp iteration too. The program was still fantastic. So whatever they are doing to make this work seems able to function despite his crankery. With a five year track record, I’m not too worried about this factor.
All but 2 of the papers listed on Manifund as coming from AISC projects are from 2021 or earlier.
In the first link at least, there are only eight papers listed in total though. With the first camp being in 2018, it doesn’t really seem like the rate dropped much? So to the extent you believe your colleagues that the camp used to be good, I don’t think the publication record is much evidence that it isn’t anymore. Paper production apparently just does not track the effectiveness of the program much. Which doesn’t surprise me, I don’t think the rate of paper producion tracks the quality of AIS research orgs much either.
The impact assessment was commissioned by AISC, not independent. They also use the number of AI alignment researchers created as an important metric. But impact is heavy-tailed, so the better metric is value of total research produced. Because there seems to be little direct research, to estimate the impact we should count the research that AISC alums from the last two years go on to produce. Unfortunately I don’t have time to do this.
Agreed on the metric being not great, and that an independently commissioned report would be better evidence (though who would have comissioned it?). But ultimately, most of what this report is apparently doing is just asking a bunch of AIS alumni what they thought of the camp and what they were up to, these days. And then noticing that these alumni often really liked it and have apparently gone on to form a significant fraction of the ecosystem. And I don’t think they even caught everyone. IIRC our AISC follow-up LTFF grant wasn’t part of the spreadsheets until I wrote Remmelt that it wasn’t there.
I am not surprised by this. Like you, my experience is that most of my current colleagues who were part of AISC tell me it was really good. The survey is just asking around and noticing the same.
I was the private donor who gave €5K. My reaction to hearing that AISC was not getting funding was that this seemed insane. The iteration I was in two years ago was fantastic for me, and the research project I got started on there is basically still continuing at Apollo now. Without AISC, I think there’s a good chance I would never have become an AI notkilleveryoneism researcher.
It feels like a very large number of people I meet in AIS today got their start in one AISC iteration or another, and many of them seem to sing its praises. I think 4⁄6 people currently on our interp team were part of one of the camps. I am not aware of any other current training program that seems to me like it would realistically replace AISC’s role, though I admittedly haven’t looked into all of them. I haven’t paid much attention to the iteration that happened in 2023, but I happen to know a bunch of people who are in the current iteration and think trying to run a training program for them is an obviously good idea.
I think MATS and co. are still way too tiny to serve all the ecosystem’s needs, and under those circumstances, shutting down a training program with an excellent five year track record seems like an even more terrible idea than usual. On top of that, the research lead structure they’ve been trying out for this camp and the last one seems to me like it might have some chance of being actually scalable. I haven’t spend much time looking at the projects for the current iteration yet, but from very brief surface exposure they didn’t seem any worse on average than the ones in my iteration. Which impressed and surprised me, because these projects were not proposed by established mentors like the ones in my iteration were. A far larger AISC wouldn’t be able to replace what a program like MATS does, but it might be able to do what AISC6 did for me, and do it for far more people than anything structured like MATS realistically ever could.
On a more meta point, I have honestly not been all that impressed with the average competency of the AIS funding ecosystem. I don’t think it not funding a project is particularly strong evidence that the project is a bad idea.
Me and Remmelt have different beliefs about AI risk, which is why the last AISC was split into two streams. Each of us are allowed to independently accept project into our own stream.
Remmelt believes that AGI alignment is impossible, i.e. there is no way to make AGI safe. Exactly why Remmelt believes this is complicated, and something I my self is still trying to understand, however this is actually not very important for AISC.
The consequence of this for this on AISC is that Remmelt is only interested in project that aims to stop AI progress.
I still think that alignment is probably technically possible, but I’m not sure. I also believe that even if alignment is possible, we need more time to solve it. Therefore, I see project that aim to stop or slow down AI progress as good, as long as there are not too large adverse side-effect. Therefore, I’m happy to have Remmelt and the projects in his stream as part of AISC. Not to mention that me an Remmelt work really well together, despite or different beliefs.
If you check our website, you’ll also notice that most of the projects are in my stream. I’ve been accepting any project as long as the there is a reasonable plan, there is a theory of change under some reasonable and self consistent assumptions, and the downside risk is not too large.
I’ve bounced around a lot in AI safety, trying out different ideas, stared more research projects than I finished, which has given me a wide view of different perspectives. I’ve updated many times in many directions, which have left me with a wide uncertainty as to what perspective is correct. This is reflected in what projects I accept to AISC. I believe in a “lets try everything” approach.
At this point, someone might think: If AISC is not filtering the project more than just “seems worth a try”, then how do AISC make sure not to waist participants time on bad projects.
Our participants are adults, and we treat them as such. We do our best to present what AISC is, and what to expect, and then let people decide for themselves if it seems like something worth their time.
We also require research leads to do the same. I.e. the project plan has to provide enough information for potential participants to judge if this is something they want to join.
I believe there is a significant chance that the solution to alignment is something no-one has though of yet. I also believe that the only way to do intellectual exploration is to let people follow their own ideas, and avoid top down curation.
The only thing I filter hard for in my stream is that the research lead actually need to have a theory of change. They need to have actually though about AI risk, and why their plan could make a difference. I had this conversation with every research lead in my stream.
We had one person last AISC who said that they regretted joining AISC, because they could have learned more from spending that time on other things. I take that feedback seriously. But on the other hand, I’ve regularly meet alumni who tell me how useful AISC was for them, which convinces me AISC is clearly very net positive.
However, if we where not understaffed (due to being underfunded), we could do more to support the research leads to make better projects.
But on the other hand, I’ve regularly meet alumni who tell me how useful AISC was for them, which convinces me AISC is clearly very net positive.
Naive question, but does AISC have enough of such past alumni that you could meet your current funding need by asking them for support? It seems like they’d be in the best position to evaluate the program and know that it’s worth funding.
Maybe I’m being cynical, but I’d give >30% that funders have declined to fund AI Safety Camp in its current form for some good reason. Has anyone written the case against?
Oli’s comment so people don’t need to click through
I thought some about the AI Safety camp for the LTFF. I mostly evaluated the research leads they listed and the resulting teams directly, for the upcoming program (which was I think the virtual one in 2023).
I felt unexcited about almost all the research directions and research leads, and the camp seemed like it was aspiring to be more focused on the research lead structure than past camps, which increased the weight I was assigning to my evaluation of those research directions. I considered for a while to fund just the small fraction of research lead teams I was excited about, but it was only a quite small fraction, and so recommended against funding it.
It did seem to me that the quality of research leads was very marketly worse by my lights than past years, so I didn’t feel comfortable just doing an outside-view on the impact of past camps (as the ARB report seems to do). I feel pretty good about the past LTFF grants to the past camps but my expectations for post-2021 camps were substantially worse than earlier camps, looking at the inputs and plans, so my expectation of the value of it substantially changed.
MATS has steadily increased in quality over the past two years, and is now more prestigious than AISC. We also have Astra, and people who go directly to residencies at OpenAI, Anthropic, etc. One should expect that AISC doesn’t attract the best talent.
There is so much wrong here, I don’t even know how to start (i.e. I don’t know what the core cruxes are) but I’ll give it a try.
I AISC is not MATS because we’re not trying to be MATS.
MATS is trying to find the best people and have them mentored by the best mentors, in the best environment. This is great! I’d recommend MATS to anyone who can get in. However it’s not scalable. After MATS has taken the top talent and mentors, there are still dosens of people who can mentor and would be happy to do so, and hundreds of people who it is worth mentoring.
To believe that MATS style program is the only program worth running, you have to believe that
Only the top talent matter
MATS and similar program has perfect selection, i.e. no-one worth accepting is ever rejected.
I’m not going to argue about 1. I suspect it’s wrong, but I’m not very sure.
However, believing in 1 is not enough. You also need 2, and believing in 2 is kind of insane. I don’t know how else to put it. Sorry.
You’re absolutely correct that AISC have lower average talent. But because we have a lower bar, we get the talent that MATS and other prestigious programs are missing.
AISC is this way by design. The idea of AISC is to give as many people as we can the chance to join the AI safety effort, to try the waters, or to show the world what they can do, or to get inspiration to do something else.
And I’m not even addressing the accessibility of a part time online program. There are people who can’t join MATS and similar, because they can’t take the time to do so, but can join AISC.
Also, if you believe strongly in MATS ability to select for talent, then consider that some AISC participants go to attend MATS later. I think this fact proves my point, that AISC can support people that MATS selection proses don’t yet recognise.
If so, AISC might not make efficient use of mentor / PI time, which is a key goal of MATS and one of the reasons it’s been successful.
This is again missing the point. The deal AISC offers to our research leads, is that they provide a project and we help them find people to work with them. So far our research leads have been very happy with this arrangement.
MATS is drawing their mentors from a small pool of well known people. This means that they have to make the most out of a very scarce resource. We’re not doing that.
AISC has an open application for people interested in leading a project. This way we get research leads you’ve never heard of, and who are happy to spend time on AISC in exchange for extra hands on their projects.
One reason AISC is much more scalable than MATS is that we’re drawing from a much larger pool of “mentors”.
At this point, someone might think: So AISC has inexperienced mentors leading inexperienced participants. How does this possibly go well?
This is not a trivial question. This is a big part of what the current version of ASIC is focusing on solving. First of all, a research lead is not the same as a mentor. Research leads are welcome to provide mentorship to it’s participants, but that’s not their main role.
The research leads role is to suggest a project and formulate a project plan, and then to lead that project. This is actually much easier to do than to provide general mentorship.
A key part of this are the project plans. As part of the application proses for research leads, we require them to write down a project plan. When necessary, we help them with this.
Another key part of how AISC is successful with less experienced “mentors”, is that we require our research leads to take active part in their projects. This obviously takes up more of their time, but also makes things work better, and to a large extent makes up for the research leads being less experienced than in other programs. And as mentioned, we get lots of project leads who are happy with this arrangement.
What the participants get is learning by doing by being part of a project that at least aims to reduce AI risk.
Some of our participants comes from AI safety Fundamentals and other such courses. Other people are professionals with various skills and talent, but not yet much involvement in AI Safety. We help these people to take the step from AI safety student or AI safety concerned professional, to being someone who actually do something.
Going from just thinking and learning, to actively engaging, is a very big step, and a lot of people would not have taken that step, or taken it later, if not for AISC.
All but 2 of the papers listed on Manifund as coming from AISC projects are from 2021 or earlier. Because I’m interested in the current quality in the presence of competing programs, I looked at the two from 2022 or later: this in a second-tier journal and this in a NeurIPS workshop, with no top conference papers. I count 52 participants in the last AISC so this seems like a pretty poor rate, especially given that 2022 and 2023 cohorts (#7 and #8) could both have published by now.
[...] They also use the number of AI alignment researchers created as an important metric. But impact is heavy-tailed, so the better metric is value of total research produced. Because there seems to be little direct research, to estimate the impact we should count the research that AISC alums from the last two years go on to produce. Unfortunately I don’t have time to do this.
That list of papers is for direct research output of AISC. Many of our alumni have lots of publications not on that list.
Just looking at the direct project outputs is not a good metric for evaluating AISC since most of the value comes from the upskilling. Counting the research that AISC alumns have done since AISC, is not a bad idea, but as you say, a lot more work, I imagine this is partly why Arb chose to do it the way they did.
I agree that heavy tailed-ness in research output is an important considerations. AISC do have some very successful alumni. If we didn’t this would be a major strike against AISC. The thing I’m less certain of is to what extent these people would have succeeded without AISC. This is obviously a difficult thing to evaluate, but still worth trying.
Mostly we let Arb decide how to best to their evaluation, but I’ve specifically asked them to interview our most successful alumni to at least get these peoples estimate of the importance of AISC. The result of this will be presented in their second report.
I suggest people actually dig themselves for evidence as to whether the program is working.
The first four points you raised seem to rely on prestige or social proof. While those can be good indicators of merit, they are also gameable.
Ie.
one program can focus on ensuring they are prestigious (to attract time-strapped alignment mentors and picky grantmakers)
another program can decide not to (because they’re not willing to sacrifice other aspects they care about).
If there is one thing you can take away from Linda and I is that we do not focus on acquiring prestige. Even the name “AI Safety Camp” is not prestigious. It sounds kinda like a bootcamp. I prefer the name because it keeps away potential applicants who are in it for the social admiration or influence.
AISC might not make efficient use of mentor / PI time, which is a key goal of MATS and one of the reasons it’s been successful.
You are welcome to ask research leads of the current edition.
“Resource-efficiency: We are not competing with other programs for scarce mentor time. Instead, we prospect for thoughtful research leads who at some point could become well-recognized researchers.”
All but 2 of the papers listed on Manifund as coming from AISC projects are from 2021 or earlier… Because I’m interested in the current quality in the presence of competing programs, I looked at the two from 2022 or later: this in a second-tier journal and this in a NeurIPS workshop, with no top conference papers.
We also do not focus on getting participants to submit papers to highly selective journals or ML conferences (though not necessarily highly selective for quality of research with regards to preventing AI-induced extinction).
AI Safety Camp is about enabling researchers that are still on the periphery of the community to learn by doing and test their fit for roles in which they can help ensure future AI are safe.
So the way to see the papers that were published is what happened after organisers did not optimise for the publication of papers, and some came out anyway.
Most groundbreaking AI Safety research that people now deem valuable was not originally published in a peer-reviewed journal. I do not think we should aim for prestigious venues now.
I would consider published papers as part of a ‘sanity check’ for evaluating editions after the fact.
If the relative number of (weighted) published papers, received grants, and org positions would have gone down for later editions, that would have been concerning. You are welcome to do your own analysis here.
Because there seems to be little direct research…
What do you mean with this claim?
If you mean research outputs, I would suggest not just focussing on peer-reviewed papers but include LessWrong/AF posts as well. Here is an overview of ~50 research outputs from past camps.
Again, AI Safety Camp acts as a training program for people who are often new to the community. The program is not like MATS in that sense.
It is relevant to consider the quality of research thinking coming out of the camp. If you or someone else had the time to look through some of those posts, I’m curious to get your sense.
Why does the founder, Remmelt Ellen, keep posting things described as…
For the record, I’m at best a co-founder.
Linda was the first camp’s initiator. Credit to her.
Now on to your point:
If you clicked through Paul’s somewhat hyperbolic comment of “the entire scientific community would probably consider this writing to be crankery” and consider my response, what are your thoughts on whether that response is reasonable or not?
Ie. consider whether the response is relevant, soundly premised, and consistently reasoned.
If you really want social proof, consider that the ex-Pentagon engineer whom Paul was reacting to got $170K in funding from SFF and has now discussed the argument in-depth for 6 hours with a long-time research collaborator (Anders Sandberg). If you would ask Anders about the post about causality limits described by a commenter as “stream of consciousness”, Anders could explain to you what the author intended to convey.
Perhaps dismissing a new relevant argument out of hand, particularly if it does not match intuitions and motivations common to our community, is not the best move?
Acknowledging here: I should not have shared some of those linkposts because they were not polished enough and did not do a good job at guiding people through the reasoning about fundamental controllability limits and substrate-needs convergence. That ended up causing more friction. My bad.
--> Edit: more here
The impact assessment was commissioned by AISC, not independent.
This is a valid concern. I have worried about conflicts of interest.
I really wanted the evaluators at Arb to do neutral research, without us organisers getting in the way. Linda and I both emphasised this at an orienting call they invited us too.
From Arb’s side, Gavin deliberately stood back and appointed Sam Holton as the main evaluator, who has no connections with AI Safety Camp. Misha did participate in early editions of the camp though.
All in, this is enough to take the report with a grain of salt. Worth picking apart the analysis and looking for any unsound premises.
My comment is not an ad hominem. An ad hominem attack would be if someone is arguing point X and you distract from X by attacking their character. I was questioning only Remmelt’s ability to distinguish good research from crankery, which is directly relevant to the job of an AISC organizer, especially because some AISC streams are about the work in question by Forrest Landry. I apologize if I was unintentionally making some broader character attack. Whether it’s obnoxious is up to you to judge.
Semantically, you could have said the same thing in far less muckrakey language - ‘Remmelt has posted widely criticised work’, for example. Yes, that’s less specific, but it’s also more important—the idea that someone should be discredited because someone said a bad thing about something they wrote is disturbingly bad epistemics.
Etymologically, your definition of an ad hominem is wrong—it can also be about attacking their circumstances. Obviously circumstances can have evidential importance, but I think it’s also poor epistemics to describe them without drawing the explicit line of inference to your conclusion—e.g. ‘Remmelt has posted and linked to widely criticised and controversial work. Will that that be represented on the curriculum?’
If you think the curriculum was or is likely to be bad, you should say that—and preferably give some specific reasons why, beyond the inferences above. Maybe just extend a tiny principle of charity when thinking about how to people who are doing their best to make the world better, and have a track record of decent work which they’ve almost certainly done for far less pay and job security than you have.
You can do all of the above and still raise at least as strong and clear questions about whether the project deserves funding.
Thomas’ comment was not ad hominem. But I personally think it is somewhat problematic.
Arepo’s counterresponse indicates why.
Collecting a pile of commenters’ negative responses to someone’s writings is not a reliable way to judge whether someone’s writing makes sense or not.
The reason being that alternative hypotheses exist that you would need to test against:
Maybe the argument is hard to convey? Maybe the author did a bad job at conveying the argument?
Maybe the writing is unpopular, for reasons unrelated to whether premises are sound and the logic holds up?
Maybe commenters did not spent much time considering the writing (perhaps it’s hard to interpret, or they disfavour the conclusion?), but used already cached mental frameworks to come to an opinion?
Maybe, for reasons like the above, this is an area where you cannot rely on the “wisdom of the crowd”?
If you have not tested against those alternative hypotheses, you are conveying more of a social intuition (others don’t seem to like this and I guess they have reasons) than a grounded judgement about whether someone else is reasoning correctly.
Crossposted from LessWrong.
Maybe I’m being cynical, but I’d give >30% that funders have declined to fund AI Safety Camp in its current form for some good reason. Has anyone written the case against? I know that AISC used to be good by talking to various colleagues, but I have no particular reason to believe in its current quality.
MATS has steadily increased in quality over the past two years, and is now more prestigious than AISC. We also have Astra, and people who go directly to residencies at OpenAI, Anthropic, etc. One should expect that AISC doesn’t attract the best talent.
If so, AISC might not make efficient use of mentor / PI time, which is a key goal of MATS and one of the reasons it’s been successful.
Why does the founder, Remmelt Ellen, keep linkposting writing by Forrest Landry which I’m 90% sure is obvious crankery? It’s not just my opinion; Paul Christiano said “the entire scientific community would probably consider this writing to be crankery”, one post was so obviously flawed it gets −46 karma, and generally the community response has been extremely negative. Some AISC work is directly about the content in question. This seems like a concern especially given the philosophical/conceptual focus of AISC projects, and the historical difficulty in choosing useful AI alignment directions without empirical grounding. [Edit: To clarify, this is not meant to be a character attack. I am concerned that Remmelt does not have the skill of distinguishing crankery from good research, even if he has substantially contributed to AISC’s success in the past.]
All but 2 of the papers listed on Manifund as coming from AISC projects are from 2021 or earlier. Because I’m interested in the current quality in the presence of competing programs, I looked at the two from 2022 or later: this in a second-tier journal and this in a NeurIPS workshop, with no top conference papers. I count 52 participants in the last AISC so this seems like a pretty poor rate, especially given that 2022 and 2023 cohorts (#7 and #8) could both have published by now. (though see this reply from Linda on why most of AISC’s impact is from upskilling)
The impact assessment was commissioned by AISC, not independent. They also use the number of AI alignment researchers created as an important metric. But impact is heavy-tailed, so the better metric is value of total research produced. Because there seems to be little direct research, to estimate the impact we should count the research that AISC alums from the last two years go on to produce. Unfortunately I don’t have time to do this.
Crossposted from LessWrong.
AISC isn’t trying to do what MATS does. Anecdotal, but for me, MATS could not have replaced AISC (spring 2022 iteration). It’s also, as I understand it, trying to have a structure that works without established mentors, since that’s one of the large bottlenecks constraining the training pipeline.
Also, did most of the past camps ever have lots of established mentors? I thought it was just the one in 2022 that had a lot? So whatever factors made all the past AISCs work and have participants sing their praises could just still be there.
He was posting cranky technical stuff during my camp iteration too. The program was still fantastic. So whatever they are doing to make this work seems able to function despite his crankery. With a five year track record, I’m not too worried about this factor.
In the first link at least, there are only eight papers listed in total though. With the first camp being in 2018, it doesn’t really seem like the rate dropped much? So to the extent you believe your colleagues that the camp used to be good, I don’t think the publication record is much evidence that it isn’t anymore. Paper production apparently just does not track the effectiveness of the program much. Which doesn’t surprise me, I don’t think the rate of paper producion tracks the quality of AIS research orgs much either.
Agreed on the metric being not great, and that an independently commissioned report would be better evidence (though who would have comissioned it?). But ultimately, most of what this report is apparently doing is just asking a bunch of AIS alumni what they thought of the camp and what they were up to, these days. And then noticing that these alumni often really liked it and have apparently gone on to form a significant fraction of the ecosystem. And I don’t think they even caught everyone. IIRC our AISC follow-up LTFF grant wasn’t part of the spreadsheets until I wrote Remmelt that it wasn’t there.
I am not surprised by this. Like you, my experience is that most of my current colleagues who were part of AISC tell me it was really good. The survey is just asking around and noticing the same.
I was the private donor who gave €5K. My reaction to hearing that AISC was not getting funding was that this seemed insane. The iteration I was in two years ago was fantastic for me, and the research project I got started on there is basically still continuing at Apollo now. Without AISC, I think there’s a good chance I would never have become an AI notkilleveryoneism researcher.
It feels like a very large number of people I meet in AIS today got their start in one AISC iteration or another, and many of them seem to sing its praises. I think 4⁄6 people currently on our interp team were part of one of the camps. I am not aware of any other current training program that seems to me like it would realistically replace AISC’s role, though I admittedly haven’t looked into all of them. I haven’t paid much attention to the iteration that happened in 2023, but I happen to know a bunch of people who are in the current iteration and think trying to run a training program for them is an obviously good idea.
I think MATS and co. are still way too tiny to serve all the ecosystem’s needs, and under those circumstances, shutting down a training program with an excellent five year track record seems like an even more terrible idea than usual. On top of that, the research lead structure they’ve been trying out for this camp and the last one seems to me like it might have some chance of being actually scalable. I haven’t spend much time looking at the projects for the current iteration yet, but from very brief surface exposure they didn’t seem any worse on average than the ones in my iteration. Which impressed and surprised me, because these projects were not proposed by established mentors like the ones in my iteration were. A far larger AISC wouldn’t be able to replace what a program like MATS does, but it might be able to do what AISC6 did for me, and do it for far more people than anything structured like MATS realistically ever could.
On a more meta point, I have honestly not been all that impressed with the average competency of the AIS funding ecosystem. I don’t think it not funding a project is particularly strong evidence that the project is a bad idea.
I see your concern.
Me and Remmelt have different beliefs about AI risk, which is why the last AISC was split into two streams. Each of us are allowed to independently accept project into our own stream.
Remmelt believes that AGI alignment is impossible, i.e. there is no way to make AGI safe. Exactly why Remmelt believes this is complicated, and something I my self is still trying to understand, however this is actually not very important for AISC.
The consequence of this for this on AISC is that Remmelt is only interested in project that aims to stop AI progress.
I still think that alignment is probably technically possible, but I’m not sure. I also believe that even if alignment is possible, we need more time to solve it. Therefore, I see project that aim to stop or slow down AI progress as good, as long as there are not too large adverse side-effect. Therefore, I’m happy to have Remmelt and the projects in his stream as part of AISC. Not to mention that me an Remmelt work really well together, despite or different beliefs.
If you check our website, you’ll also notice that most of the projects are in my stream. I’ve been accepting any project as long as the there is a reasonable plan, there is a theory of change under some reasonable and self consistent assumptions, and the downside risk is not too large.
I’ve bounced around a lot in AI safety, trying out different ideas, stared more research projects than I finished, which has given me a wide view of different perspectives. I’ve updated many times in many directions, which have left me with a wide uncertainty as to what perspective is correct. This is reflected in what projects I accept to AISC. I believe in a “lets try everything” approach.
At this point, someone might think: If AISC is not filtering the project more than just “seems worth a try”, then how do AISC make sure not to waist participants time on bad projects.
Our participants are adults, and we treat them as such. We do our best to present what AISC is, and what to expect, and then let people decide for themselves if it seems like something worth their time.
We also require research leads to do the same. I.e. the project plan has to provide enough information for potential participants to judge if this is something they want to join.
I believe there is a significant chance that the solution to alignment is something no-one has though of yet. I also believe that the only way to do intellectual exploration is to let people follow their own ideas, and avoid top down curation.
The only thing I filter hard for in my stream is that the research lead actually need to have a theory of change. They need to have actually though about AI risk, and why their plan could make a difference. I had this conversation with every research lead in my stream.
We had one person last AISC who said that they regretted joining AISC, because they could have learned more from spending that time on other things. I take that feedback seriously. But on the other hand, I’ve regularly meet alumni who tell me how useful AISC was for them, which convinces me AISC is clearly very net positive.
However, if we where not understaffed (due to being underfunded), we could do more to support the research leads to make better projects.
Naive question, but does AISC have enough of such past alumni that you could meet your current funding need by asking them for support? It seems like they’d be in the best position to evaluate the program and know that it’s worth funding.
We have reached out to them and gotten some donations.
The “Do Not Build Uncontrollable AI” area is meant for anyone to join who have this concern.
The purpose of this area is to contribute to restricting corporations from recklessly scaling the training and uses of ML models.
I want the area to be open for contributors who think that:
we’re not on track to solving safe control of AGI; and/or
there are fundamental limits to the controllability of AGI, and unfortunately AGI cannot be kept safe over the long term; and/or
corporations are causing increasing harms in how they scale uses of AI models.
After thinking about this over three years, I now think 1.-3. are all true.
I would love more people who hold any of these views to collaborate thoughtfully across the board!
To keep communication open here, here is Oliver Habryka’s LessWrong comment.
Oli’s comment so people don’t need to click through
Here are some evaluations not commissioned by us
Takeaways from a survey on AI alignment resources — EA Forum (effectivealtruism.org)
Thoughts on AI Safety Camp — EA Forum (effectivealtruism.org)
If you have suggestions for how AISC can get more people to do more independent evaluations, please let me know.
There is so much wrong here, I don’t even know how to start (i.e. I don’t know what the core cruxes are) but I’ll give it a try.
I AISC is not MATS because we’re not trying to be MATS.
MATS is trying to find the best people and have them mentored by the best mentors, in the best environment. This is great! I’d recommend MATS to anyone who can get in. However it’s not scalable. After MATS has taken the top talent and mentors, there are still dosens of people who can mentor and would be happy to do so, and hundreds of people who it is worth mentoring.
To believe that MATS style program is the only program worth running, you have to believe that
Only the top talent matter
MATS and similar program has perfect selection, i.e. no-one worth accepting is ever rejected.
I’m not going to argue about 1. I suspect it’s wrong, but I’m not very sure.
However, believing in 1 is not enough. You also need 2, and believing in 2 is kind of insane. I don’t know how else to put it. Sorry.
You’re absolutely correct that AISC have lower average talent. But because we have a lower bar, we get the talent that MATS and other prestigious programs are missing.
AISC is this way by design. The idea of AISC is to give as many people as we can the chance to join the AI safety effort, to try the waters, or to show the world what they can do, or to get inspiration to do something else.
And I’m not even addressing the accessibility of a part time online program. There are people who can’t join MATS and similar, because they can’t take the time to do so, but can join AISC.
Also, if you believe strongly in MATS ability to select for talent, then consider that some AISC participants go to attend MATS later. I think this fact proves my point, that AISC can support people that MATS selection proses don’t yet recognise.
This is again missing the point. The deal AISC offers to our research leads, is that they provide a project and we help them find people to work with them. So far our research leads have been very happy with this arrangement.
MATS is drawing their mentors from a small pool of well known people. This means that they have to make the most out of a very scarce resource. We’re not doing that.
AISC has an open application for people interested in leading a project. This way we get research leads you’ve never heard of, and who are happy to spend time on AISC in exchange for extra hands on their projects.
One reason AISC is much more scalable than MATS is that we’re drawing from a much larger pool of “mentors”.
At this point, someone might think: So AISC has inexperienced mentors leading inexperienced participants. How does this possibly go well?
This is not a trivial question. This is a big part of what the current version of ASIC is focusing on solving. First of all, a research lead is not the same as a mentor. Research leads are welcome to provide mentorship to it’s participants, but that’s not their main role.
The research leads role is to suggest a project and formulate a project plan, and then to lead that project. This is actually much easier to do than to provide general mentorship.
A key part of this are the project plans. As part of the application proses for research leads, we require them to write down a project plan. When necessary, we help them with this.
Another key part of how AISC is successful with less experienced “mentors”, is that we require our research leads to take active part in their projects. This obviously takes up more of their time, but also makes things work better, and to a large extent makes up for the research leads being less experienced than in other programs. And as mentioned, we get lots of project leads who are happy with this arrangement.
What the participants get is learning by doing by being part of a project that at least aims to reduce AI risk.
Some of our participants comes from AI safety Fundamentals and other such courses. Other people are professionals with various skills and talent, but not yet much involvement in AI Safety. We help these people to take the step from AI safety student or AI safety concerned professional, to being someone who actually do something.
Going from just thinking and learning, to actively engaging, is a very big step, and a lot of people would not have taken that step, or taken it later, if not for AISC.
That list of papers is for direct research output of AISC. Many of our alumni have lots of publications not on that list.
For example, I looked up Marius Hobbhahn—Google Scholar
Just looking at the direct project outputs is not a good metric for evaluating AISC since most of the value comes from the upskilling. Counting the research that AISC alumns have done since AISC, is not a bad idea, but as you say, a lot more work, I imagine this is partly why Arb chose to do it the way they did.
I agree that heavy tailed-ness in research output is an important considerations. AISC do have some very successful alumni. If we didn’t this would be a major strike against AISC. The thing I’m less certain of is to what extent these people would have succeeded without AISC. This is obviously a difficult thing to evaluate, but still worth trying.
Mostly we let Arb decide how to best to their evaluation, but I’ve specifically asked them to interview our most successful alumni to at least get these peoples estimate of the importance of AISC. The result of this will be presented in their second report.
Glad you raised these concerns!
I suggest people actually dig themselves for evidence as to whether the program is working.
The first four points you raised seem to rely on prestige or social proof. While those can be good indicators of merit, they are also gameable.
Ie.
one program can focus on ensuring they are prestigious (to attract time-strapped alignment mentors and picky grantmakers)
another program can decide not to (because they’re not willing to sacrifice other aspects they care about).
If there is one thing you can take away from Linda and I is that we do not focus on acquiring prestige. Even the name “AI Safety Camp” is not prestigious. It sounds kinda like a bootcamp. I prefer the name because it keeps away potential applicants who are in it for the social admiration or influence.
You are welcome to ask research leads of the current edition.
Note from the Manifund post:
“Resource-efficiency: We are not competing with other programs for scarce mentor time. Instead, we prospect for thoughtful research leads who at some point could become well-recognized researchers.”
We also do not focus on getting participants to submit papers to highly selective journals or ML conferences (though not necessarily highly selective for quality of research with regards to preventing AI-induced extinction).
AI Safety Camp is about enabling researchers that are still on the periphery of the community to learn by doing and test their fit for roles in which they can help ensure future AI are safe.
So the way to see the papers that were published is what happened after organisers did not optimise for the publication of papers, and some came out anyway.
Most groundbreaking AI Safety research that people now deem valuable was not originally published in a peer-reviewed journal. I do not think we should aim for prestigious venues now.
I would consider published papers as part of a ‘sanity check’ for evaluating editions after the fact. If the relative number of (weighted) published papers, received grants, and org positions would have gone down for later editions, that would have been concerning. You are welcome to do your own analysis here.
What do you mean with this claim?
If you mean research outputs, I would suggest not just focussing on peer-reviewed papers but include LessWrong/AF posts as well. Here is an overview of ~50 research outputs from past camps.
Again, AI Safety Camp acts as a training program for people who are often new to the community. The program is not like MATS in that sense.
It is relevant to consider the quality of research thinking coming out of the camp. If you or someone else had the time to look through some of those posts, I’m curious to get your sense.
For the record, I’m at best a co-founder. Linda was the first camp’s initiator. Credit to her.
Now on to your point:
If you clicked through Paul’s somewhat hyperbolic comment of “the entire scientific community would probably consider this writing to be crankery” and consider my response, what are your thoughts on whether that response is reasonable or not? Ie. consider whether the response is relevant, soundly premised, and consistently reasoned.
If you really want social proof, consider that the ex-Pentagon engineer whom Paul was reacting to got $170K in funding from SFF and has now discussed the argument in-depth for 6 hours with a long-time research collaborator (Anders Sandberg). If you would ask Anders about the post about causality limits described by a commenter as “stream of consciousness”, Anders could explain to you what the author intended to convey.
Perhaps dismissing a new relevant argument out of hand, particularly if it does not match intuitions and motivations common to our community, is not the best move?
Acknowledging here: I should not have shared some of those linkposts because they were not polished enough and did not do a good job at guiding people through the reasoning about fundamental controllability limits and substrate-needs convergence. That ended up causing more friction. My bad. --> Edit: more here
This is a valid concern. I have worried about conflicts of interest.
I really wanted the evaluators at Arb to do neutral research, without us organisers getting in the way. Linda and I both emphasised this at an orienting call they invited us too.
From Arb’s side, Gavin deliberately stood back and appointed Sam Holton as the main evaluator, who has no connections with AI Safety Camp. Misha did participate in early editions of the camp though.
All in, this is enough to take the report with a grain of salt. Worth picking apart the analysis and looking for any unsound premises.
Why does this commenter post things described as ‘very, very elitist’, work for an organisation described as having ‘wasted’ $46million on ‘nothing of any obvious use’, and post comments so obviously callous and uncharitable they’ve been described as ‘an obnoxious ad hominem with no place on this forum’?
Your comment, incidentally, is an obnoxious ad hominem with no place on this forum.
My comment is not an ad hominem. An ad hominem attack would be if someone is arguing point X and you distract from X by attacking their character. I was questioning only Remmelt’s ability to distinguish good research from crankery, which is directly relevant to the job of an AISC organizer, especially because some AISC streams are about the work in question by Forrest Landry. I apologize if I was unintentionally making some broader character attack. Whether it’s obnoxious is up to you to judge.
Semantically, you could have said the same thing in far less muckrakey language - ‘Remmelt has posted widely criticised work’, for example. Yes, that’s less specific, but it’s also more important—the idea that someone should be discredited because someone said a bad thing about something they wrote is disturbingly bad epistemics.
Etymologically, your definition of an ad hominem is wrong—it can also be about attacking their circumstances. Obviously circumstances can have evidential importance, but I think it’s also poor epistemics to describe them without drawing the explicit line of inference to your conclusion—e.g. ‘Remmelt has posted and linked to widely criticised and controversial work. Will that that be represented on the curriculum?’
If you think the curriculum was or is likely to be bad, you should say that—and preferably give some specific reasons why, beyond the inferences above. Maybe just extend a tiny principle of charity when thinking about how to people who are doing their best to make the world better, and have a track record of decent work which they’ve almost certainly done for far less pay and job security than you have.
You can do all of the above and still raise at least as strong and clear questions about whether the project deserves funding.
To lay a middle ground here:
Thomas’ comment was not ad hominem. But I personally think it is somewhat problematic.
Arepo’s counterresponse indicates why.
Collecting a pile of commenters’ negative responses to someone’s writings is not a reliable way to judge whether someone’s writing makes sense or not.
The reason being that alternative hypotheses exist that you would need to test against:
Maybe the argument is hard to convey? Maybe the author did a bad job at conveying the argument?
Maybe the writing is unpopular, for reasons unrelated to whether premises are sound and the logic holds up?
Maybe commenters did not spent much time considering the writing (perhaps it’s hard to interpret, or they disfavour the conclusion?), but used already cached mental frameworks to come to an opinion?
Maybe, for reasons like the above, this is an area where you cannot rely on the “wisdom of the crowd”?
If you have not tested against those alternative hypotheses, you are conveying more of a social intuition (others don’t seem to like this and I guess they have reasons) than a grounded judgement about whether someone else is reasoning correctly.
For what it’s worth, I think it was good that Thomas brought this up so that we could respond.
Also see further discussion on LessWrong here and here.