Posts tagged community
Quick takes
Popular comments
Recent discussion
I'm launching AI Lab Watch. I collected actions for frontier AI labs to improve AI safety, then evaluated some frontier labs accordingly.
It's a collection of information on what labs should do and what labs are doing. It also has some adjacent resources, including a list...
"And as I said, I expect such a project to appear soon."
I dont know whether to read this as "Zach has some inside information that gives him high confidence it will exist" or "Zach is doing wishful thinking" or something else!
NOTE: This post was updated to include two additional models which meet the criteria for being considered Open Source AI.
As advanced machine learning systems become increasingly widespread, the question of how to make them safe is also gaining attention. Within this...
The post has now been updated appropriately, please let me know if you don't think the modifications were sufficient so that I can fix them appropriately.
Now to make the changes on the other platforms!
This announcement was written by Toby Tremlett, but don’t worry, I won’t answer the questions for Lewis.
Lewis Bollard, Program Director of Farm Animal Welfare at Open Philanthropy, will be holding an AMA on Wednesday 8th of May. Put all your questions for him on this thread...
This post was partly inspired by, and shares some themes with, this Joe Carlsmith post. My post (unsurprisingly) expresses fewer concepts with less clarity and resonance, but is hopefully of some value regardless.
Content warning: description of animal death.
I live in a ...
I've got moths in my flat right now and this post made me take solving this more seriously. Thank you!
- SoGive works with major donors.
- As part of our work, we meet with several (10-30 per year) charities, generally ones recommended by evaluators we trust, or (occasionally) recommended by our own research.
- We learn a lot through these conversations. This suggests that we might want to publish our call notes so that others can also learn about the charities we speak with.
- Given that we take notes during the calls anyway, it might seem that it would be low cost for us to simply publish those. This would be deceptive.
- There is a non-trivial time cost for us, partly because documents which are published are held to a higher standard than those which are purely internal, but mostly because of our relationship with the charities. We want them to feel confident that they can speak openly with us. This means not only an extra step in the process (ie sharing a draft with the organisation
Summary
- Where there’s overfishing, reducing fishing pressure or harvest rates — roughly the share of the population or biomass caught in a fishery per fishing period — actually allows more animals to be caught in the long run.
- Sustainable fishery management policies
Just the arguments in the summary are really solid.[1] And while I wasn't considering supporting sustainability in fishing anyway, I now believe it's more urgent to culturally/semiotically/associatively separate between welfare and some strands of "environmentalism". Thanks!
Alas, I don't predict I will work anywhere where this update becomes pivotal to my actions, but my practically relevant takeaway is: I will reproduce the arguments from this post (and/or link it) in contexts where people are discussing conjunctions/disjunctions between environmenta...
Trump recently said in an interview (https://time.com/6972973/biden-trump-bird-flu-covid/) that he would seek to disband the White House office for pandemic preparedness. Given that he usually doesn't give specifics on his policy positions, this seems like something he ...
The full quote suggests this is because he classifies Operation Warp Speed (reactive, targeted) as very different from the Office (wasteful, impossible to predict what you'll need, didn't work last time). I would classify this as a disagreement about means rather than ends.
One last question, Mr. President, because I know that your time is limited, and I appreciate your generosity. We have just reached the four-year anniversary of the COVID pandemic. One of your historic accomplishments was Operation Warp Speed. If we were to have another pandemic, would you take the same actions to manufacture and distribute a vaccine and get it in the arms of Americans as quickly as possible?
Trump: I did a phenomenal job. I appreciate the way you worded that question. So I have a very important Democrat friend, who probably votes for me, but I'm not 100% sure, because he's a serious Democrat, and he asked me about it. He said Operation Warp Speed was one of the greatest achievements in the history of government. What you did was incredible, the speed of it, and the, you know, it was supposed to take anywhere from five to 12 years, the whole thing. Not only that: the ventilators, the therapeutics, Regeneron and other things. I mean Regeneron was incredible. But therapeutics—everything. The overall—Operation Warp Speed, and you never talk about it. Democrats talk about it as if it’s the greatest achievement. So I don’t talk about it. I let others talk about it.
You know, you have strong opinions both ways on the vaccines. It's interesting. The Democrats love the vaccine. The Democrats. Only reason I don’t take credit for it. The Republicans, in many cases, don’t, although many of them got it, I can tell you. It’s very interesting. Some of the ones who talk the most. I said, “Well, you didn’t have it did you?” Well, actually he did, but you know, et cetera.
But Democrats think it’s an incredible, incredible achievement, and they wish they could take credit for it, and Republicans don’t. I don't bring it up. All I do is just, I do the right thing. And we've gotten actually a lot of credit for Operation Warp Speed. And the power and the speed was incredible. And don’t forget, when I said, nobody had any idea what this was. You know, we’re two and a half years, almost three years, nobody ever. Everybody thought of a pandemic as an ancient problem. No longer a modern problem, right? You know, you don't think of that? You hear about 1917 in Europe and all. You didn’t think that could happen. You learned if you could. But nobody saw that coming and we took over, and I’m not blaming the past administrations at all, because again, nobody saw it coming. But the cupboards were bare.
We had no gowns, we had no masks. We had no goggles, we had no medicines. We had no ventilators. We had nothing. The cupboards were totally bare. And I energized the country like nobody’s ever energized our country. A lot of people give us credit for that. Unfortunately, they’re mostly Democrats that give me the credit.
Well, sir, would you do the same thing again to get vaccines in the arms of Americans as quickly as possible, if it happened again in the next four years?
Trump: Well, there are the variations of it. I mean, you know, we also learned when that first came out, nobody had any idea what this was, this was something that nobody heard of. At that time, they didn’t call it Covid. They called it various names. Somehow they settled on Covid. It was the China virus, various other names.
But when this came along, nobody had any idea. All they knew was dust coming in from China. And there were bad things happening in China around Wuhan. You know, I predicted. I think you'd know this, but I was very strong on saying that this came from Wuhan. And it came from the Wuhan labs. And I said that from day one. Because I saw things that led me to believe that, very strongly led me to believe that. But I was right on that. A lot of people say that now that Trump really did get it right. A lot of people said, “Oh, it came from caves, or it came from other countries.” China was trying to convince people that it came from Italy and France, you know, first Italy, then France. I said, “No, it came from China, and it came from the Wuhan labs.” And that's where it ended up coming from. So you know, and I said that very early. I never said anything else actually. But I've been given a lot of credit for Operation Warp Speed. But most of that credit has come from Democrats. And I think a big portion of Republicans agree with it, too. But a lot of them don't want to say it. They don't want to talk about it.
So last follow-up: The Biden Administration created the Office of Pandemic Preparedness and Response Policy, a permanent office in the executive branch tasked with preparing for epidemics that have not yet emerged. You disbanded a similar office in 2018 that Obama had created. Would you disband Biden's office, too?
Trump: Well, he wants to spend a lot of money on something that you don't know if it's gonna be 100 years or 50 years or 25 years. And it's just a way of giving out pork. And, yeah, I probably would, because I think we've learned a lot and we can mobilize, you know, we can mobilize. A lot of the things that you do and a lot of the equipment that you buy is obsolete when you get hit with something. And as far as medicines, you know, these medicines are very different depending on what strains, depending on what type of flu or virus it may be. You know, things change so much. So, yeah, I think I would. It doesn't mean that we're not watching out for it all the time. But it's very hard to predict what's coming because there are a lot of variations of these pandemics. I mean, the variations are incredible, if you look at it. But we did a great job with the therapeutics. And, again, these therapeutics were specific to this, not for something else. So, no, I think it's just another—I think it sounds good politically, but I think it's a very expensive solution to something that won't work. You have to move quickly when you see it happening.
I'd say it's 50/50 but sure. And while politics is discouraged, I don't think that your thing is really what's being discouraged.
A crucial consideration in assessing the risks of advanced AI is the moral value we place on "unaligned" AIs—systems that do not share human preferences—which could emerge if we fail to make enough progress on technical alignment.
In this post I'll consider three potential...
Here are a few (long, but high-level) comments I have before responding to a few specific points that I still disagree with:
- I agree there are some weak reasons to think that humans are likely to be more utilitarian on average than unaligned AIs, for basically the reasons you talk about in your comment (I won't express individual agreement with all the points you gave that I agree with, but you should know that I agree with many of them).
However, I do not yet see any strong reasons supporting your view. (The main argument seems to be: AIs will be diff
There have been multiple occasions where I've copy and pasted email threads into an LLM and asked it things like:
I really want an email plugin that basically brute forces rationality INTO email conversations.
Tangentially - I wonder if LLMs can reliably convert peoples claims into a % through sentiment analysis? This would be useful for Forecasters I believe (and rationality in general)