Data scientist working on AI forecasting through Epoch and the Stanford AI Index. GWWC pledge member since 2017. Formerly social chair at Harvard Effective Altruism, facilitator for Arete Fellowship, and founder of the DC Slate Star Codex meetup.
Robi Rahman
Karma: 1,080
What are some EA/LW/etc coworking spaces that could accommodate ~10 people for ~5 days? I’m aware of Constellation and Lighthaven (Berkeley, CA), HAIST and MAIA (Cambridge, MA), Wytham Abbey and Trajan House (Oxford, UK), CEEALAR (Blackpool, UK), LEAH and LISA (London, UK), Epistea and Fixed Point (Prague, Czechia). Are there any others?
What are the stringent and permissive criteria for judging that someone has heard of EA?
They keep saying they’re working on a response. It’s probably around 500 pages by now.
Announcing Epoch’s newly expanded Parameters, Compute and Data Trends in Machine Learning database
If you’re trying to maximize computational efficiency, instead of building a Dyson sphere, shouldn’t you drop the sun into a black hole and harvest the Hawking radiation?
Upvoted your post because you made some good points, but I think your analogy between human cloning and AI training is totally wrong.
Take for example, human reproductive cloning. This is so morally abhorrent that it is not being practised anywhere in the world. There is no black market in need of a global police state to shut it down. AGI research could become equally uncool once the danger, and loss of sovereignty, it represents is sufficiently well appreciated.
There is no black market in human cloning, and no police state trying to stop it, because no one benefits very much from cloning. Cloning is just not that useful. Whereas if we stop corporate AI development for 40 years but computer hardware keeps improving, anyone can get rich by training an AGI on their gaming laptop. It would be like trying to confiscate all the drug imports in a world where everyone with a cell phone is a drug addict.
EAGxNYC 2023: Retrospective
Can you name some of the red flags to watch for? I’d also be interested in hearing who some of the bad actors are (perhaps in a DM if you don’t want them to know they’ve been spotted).
You mean this?
If so, what part of it do you object to?
This doesn’t answer your question, but: I’ve heard several people opine that “fiscal sponsorship” is a really bad name for what it entails. I work at Epoch, which is a fiscal sponsee of Rethink Priorities (and yes, RP uses the word “sponsee” for us and all their sponsees). My understanding is that we (Epoch) pay some kind of fee to RP (annual? maybe a percentage of our budget? idk), and in return, RP’s HR people handle our HR stuff and some of their ops people spend some time doing ops work for us. This is almost the complete opposite of being “fiscally sponsored” in the sense of being sponsored by and receiving money from the sponsor.
h/t @SarahPomeranz
We just finished hiring a data analyst for October. It’s possible that we’ll hire another candidate in the future, but the position is not currently taking applications.
I don’t think this speaks badly to their skill level and certainly not their potential, just that they start out in a really unfair circumstance, with a head filled with a bunch of bullshit that just needs to be thrown out as cleanly as possible, and Mearsheimer is a great way to do that.
I’m out of the loop; what’s the bullshit from high school civics class that needs to be thrown out of my head, and why is Mearsheimer unbalanced but also a good starting point?
Epoch is hiring an Associate Data Analyst
Do you have examples of laws EA orgs might ask their employees to break?
I am very happy to clarify topics around nuclear, coming from the energy industry myself.
What part of the energy industry do you work in?
For example, with $2k, I expect I could hire a pub in central London for an evening (or maybe a whole day), with perhaps around 100 people attending. So that’s $20 per person, or 1% of the cost of EAG. Would they get as much benefit from attending my event as attending EAG? No, but I’d bet they’d get more than 1% of the benefit.
Actually, I’m not sure this is right. An evening has around 1⁄10 of the networking duration of a weekend, and number of connections are proportional to time spent networking and to number of participants squared. If this is 1/10th the attendance and 1/10th the duration, you’re getting around 1/1000th the networking value for 1/1000th the cost, so not obviously better than an EAG weekend that costs $2M and has 1000 participants, maybe worse when you factor in people’s time cost of getting to the pub as a fraction of the event duration, or maybe better if they prefer not to spend a whole weekend.
Furthermore, a lion becomes more dangerous as it becomes more intelligent and capable, even if its terminal goal is not “maximize number of wildebeests eaten”.
I think you’re committing a typical mind fallacy if you think most people would benefit from reading HPMOR as much as you did.
There are lots of these! I saw several orgs that provide that service while I was working on EAGxNYC. I can look them up later and get back to you (but we’re really busy with the conference coming up so it might take a while).
Huh? What’s the lesson from FTX that would have improved the OpenAI situation?