[Comment deleted as the issue has been clarified]

I feel confused about how dangerous/costly it is to use LLMs for private documents or thoughts to assist longtermist research, in a way that may wind up in the training data for future iterations of LLMs. Some sample use cases that I'd be worried about:

• Summarizing private AI evals docs about plans to evaluate future models
• Rewrite emails on high-stakes AI gov conversations
• Generate lists of ideas for biosecurity interventions that can be helped/harmed by AI
• Scrub potentially risky/infohazard-y information from a planned public forecasting questions
• Summarize/rewrite speculations of potential near-future AI capabilities gains.

I'm worried about using LLMs for the following reasons:

1. Standard privacy concerns/leakage to dangerous (human) actors
   1. If it's possible to back out your biosecurity plans from the models, this might give ideas to terrorists/rogue gov'ts.
   2. your infohazards might leak
   3. People might (probabilistically) back out private sensitive communication, which could be embarrassing
      1. I wouldn't be surprised if care for consumer privacy at AGI labs for chatbot consumers is much lower than say for emails hosted by large tech companies
         1. I've heard rumors to this effect, also see
   4. (unlikely) your

I would like advice on writing a resume and applying to work in an effective career.  I will graduate with an economics bachelor's degree in April. I'm taking many statistics courses. I also took calculus and computer science courses. I live on the west coast of Canada and I am willing to move.

I believe I would be well suited to AI Governance but it may be better currently to find statistics/econometrics work or do survey design (to build general skills until I know more AI Governance people, or switch into a different effective cause area)

I am also o... (read more)

Hi there everyone, I'm William the Kiwi and this is my first post on EA forums. I have recently discovered AI alignment and have been reading about it for around a month. This seems like an important but terrifyingly under invested in field. I have many questions but in the interest of speed I will involve Cunningham's Law and post my current conclusions.

My AI conclusions:

1. Corrigiblity is mathematically impossible for AGI.
2. Alignment requires defining all important human values in a robust enough way that it can survive near-infinite amounts of optimisat

The table here got all messed up. Could it be fixed?

https://forum.effectivealtruism.org/posts/qZqvBLvR5hX9sEkjR/comparing-top-forecasters-and-domain-experts

Hey everyone! First time poster here, but long time advocate for effective altruism.

I've been vegan for a couple of years now, mostly to mitigate animal suffering. Recently I've been wondering how a vegetarian diet would compare in terms of suffering caused. Of course I presume veganism would be better, but by how much?

With this in mind I'm wondering is there any resources that attempt to quantify how much suffering is caused by buying various animal products?  For example dairy cows produce about 40,000 litres of milk in their lifetime, which can be ... (read more)

I'm looking for statistics on how doable it is to solve all the problems we care about. For example, I came across this: https://www.un.org/sustainabledevelopment/wp-content/uploads/2018/09/Goal-1.pdf from the UN which says extreme poverty could be sorted out in 20 years for $175 billion a year. That is actually very doable, in light of the fact of how much money can go into war (in 1945, the US spent 40% of its GDP into the war). I'm looking for more numbers like that, e.g. how much money it takes to solve X problem. 

I intend to use them for a ... (read more)

Does anyone know why Singer hasn't changed his views on infanticide and killing animals after he had become a hedonist utilitarian? As far as I know, his former views were based on the following:

a. Creation and fulfilment of new preferences is morally neutral.

b. Thwarting existing preferences is morally bad.

c. Persons have preferences about their future.

d. Non-persons don't have a sense of the future, they don't have preferences about their future either. They live in the moment.

e. Killing persons thwarts their preferences about the future.

f. Killing non-p... (read more)

Hi guys !

I posted about 2 weeks ago here asking for masters project ideas around the field of computational social choice and machine learning for ethical decision making.

To recap: I'm currently doing my master's project in design engineering at Imperial, where I need to find something impactful, implementable and innovative.

I really appreciated all the help I got on the post, however, I've hit a kind of dead end - I'm not sure I can find something within my scope with the time frame in the field I've chosen.

So now I'm asking for any project ideas which fi... (read more)

Hello everyone!  My name is Carlos. I recently realized I should be leading a life of service, instead of one where I only care about myself, and that has taken me here, to the place that is all about doing the most good.

I'm an odd guy, in that I have read some LessWrong and have been reading Slate Star Codex/Astral Codex Ten for years, but am for all intents and purposes a mystic. That shouldn't put me at odds here too much, since rationality is definitely a powerful and much needed tool in certain contexts (such as this one), it's just that it canno... (read more)

Is there a way to only show posts with ≥ 50 upvotes on the Frontpage?

Hi all,

Moya here from Darmstadt, Germany. I am a Culture-associated scientist, trans* feminist, poly, kinky, and a witch.
I got into LessWrong in 2016 and then EA 2016 or 2017, don't quite remember. :)

I went to the University of Iceland, did a Master's degree in Computer Science / Bioinformatics there, then built software for the European Space Agency, and nowadays am a freelance programmer and activist in the Seebrücke movement in Germany and other activist groups as well. I also help organize local burn events (some but not all of them being FLINTA* exclu... (read more)

First time poster here.
I am currently doing my master's degree in design engineering at Imperial College London, and I am trying to create a project proposal around the topic of computational social choice and machine learning for ethical decision making. I'm struggling to find a "design engineering" take on this - what can I do to contribute in the field as a design engineer?

In terms of prior art, I've been inspired by MIT's Moral Machine, feeding ML models of aggregate ethical decisions from people. If anyone has any ideas on a des eng angle to approach this topic, please give me some pointers! 

TIA

I have seen Sabine Hossenfelder claim that it will be very expensive to maintain superintelligent AIs. I also hear many people claiming that digital minds will use much less energy than human minds, so they will be much more numerous. Does anyone have some information or a guess on how much energy ChatGPT spends per hour per user?

It’ll be my first time at a Bay Area EA Global at the end of this month - does anyone have any tips? Any things I should definitely do?

Also if you’re interested in institutional reform you might like my blog Rules of the Game: https://connoraxiotes.substack.com/p/what-can-the-uk-government-do-to

Hello, All!

I found EA via the New Yorker article about William MacAskill.
I am the author of "Thank You For Listening". 
I listen, therefore you are. We understand and respect, therefore we are. We bring out the best in each other, therefore we thrive.
Go beyond Can Do. We Can understand, respect, and bring out the best in others, often beyond our expectations.
We know how to cooperate on roads. We can cooperate at home, at work, and in society. Teach everyone to listen (yield), check biases (blind spots), and reject ideological rage (road rage).
Bringing Out  The  Best In Humanity

Hi everyone,

I was close to becoming a statistic of someone who started reading 80,000 hours but never completed the career planning program. I am coming back now as I need some direction.

Of all the global priorities, I gravitate toward those that focus on improving physical and mental health. As someone who deals with chronic pain and is in between jobs, nothing consumes my attention more than alleviating physical and mental suffering.

I am curious if anyone in the community spends their work life thinking and working on increasing longevity, eliminating ch... (read more)

Just a warning on treating everyone as if they argue in good faith. They don’t. Émile P. Torres, aka @xriskology on Twitter doesn’t. He may say true honest things but if you find anything he says insightful check all the sources.

Émile P. Torres’s history of dishonesty and harassment An incomplete summary https://markfuentes1.substack.com/p/emile-p-torress-history-of-dishonesty

Does anyone have estimates on the cost effectiveness of trachoma prevention? It seems as though mass antibiotic administration is effective and cheap, and blindness is quite serious. However room for funding might be limited. I haven't seen it investigated by many of the organizations, but maybe I just haven't found the right report.

Hey everyone, I'm curious about the extent to which people in EA take (weak/strong) antinatalism/ negative utilitarianism seriously. I've read a bit around the topic and find some arguments more persuasive than others, but the idea that many lives are net-negative, and that even good lives might be worse than we think they are, has stuck with me. 

Based on my own mood diary, I'm leaning towards something around a 5.5/10 on a happiness scale being the neutral point, under which a life isn't worth living. 

This has made me a lot less enthusiastic about 'saving lives' for its own sake, especially those lives in countries/ regions with very poor quality of life. So I suspect that some 'life-saving' charities could be actively harmful and that we should focus way more on 'life-improving' charities/ cause areas. (There are probably very few charities that only save lives- preventing malaria/ reducing lead exposure both improves and saves lives- but we can imagine a 'pure-play life-saving charity'.)

I haven't come to any conclusions here, but the 'cost to save a life' framing, still common in EA, strikes me as probably morally invalid. I don't hear this argument mentioned much (you don't seem to get anyone actively arguing against 'saving lives'), so I'm just curious what the range of EA opinion is. 

How do folks! Stoked to have the opportunity to try and be a participant that contributes something meaningful here on the EA Forum. 

EA Forum Guidelines (and Aaron)...thank you for the guidance and encouraging me to write the bio. 

All, I'm new to the EA community. I'll hope to meet some of you soon. Please feel free to send a hello anytime. 

I see the "Commenting Guidelines". They remind me of the Simple Rules of Inquiry that I've used for many years. Are they a decent match for the spirit of this Forum?

1. Turn judgment into curiosity
2. Turn confli

Hi I’m Silas Barta. First comment here! I organize the Austin LessWrong group. I’m currently retired off of earlier investing (formerly software engineer) but am still looking for my next career to maximize my impact. I think I have a calling in either information security (esp reverse engineering) or improving the quality of explanations and introductions to technical topics.

I have donated cryptocurrency and contributed during Facebook’s Giving Tuesday, and gone to the Bay Area EA Globals in 2016 and 2017.

Qn: Where is the closest EA community base to the US? How accesible is the USA from it (US Consulate)?

Context: I am recently let go from my job while on a visa in the states. Which means I have to leave the US within the next 7 days. I would like to live somewhere close to the US where I can find community so that I don't loose momentum to do the intense work that job search needs. I tend to be really affected by the energy of where I am; I work best in cities, I tend to sleep most on a countryside.

This might also be a good resource for people who are not ... (read more)

Hi all, I'm Vlad, 35, from Romania. I've been working in software engineering for 12 years. I have a bachelor's and master's degree in Physics.

I'm here because I read "What we owe the future", after it was recommended to me by a friend.

I got the book recommended to me because I had an idea which is a little unconfortable for some people, but I think this idea is extremely important, and this friend of mine instantly classified my thoughts as "a branch of long-termism". I also think my idea is extremely relevant to this group, and I'm interested in getting feedback about it.

Context for the  idea: Long-termism is concerned about people as far into the future as possible, up to the end of the universe.

The idea: ...what if we can make it so there doesn't have to be an end? If we had a limitless source of energy, there wouldn't have to be an end. Not only that, but we could make a lot of people very happy (like billions of billions of billions .....of billions of them? a literal infinity of them even)

It sounds crazy, I realize, but my best knowledge on this topic says this:

• We know that we don't know all the laws of the universe
• Even the known laws kind of have a loop-hole in t