constructive

Karma: 560

constructive 5 Jan 2024 10:29 UTC
1 point
0 ∶ 0
in reply to: tlevin’s comment on: EU policymakers reach an agreement on the AI Act
And your update is that this process will be more globally impactful than you initially expected? Would be curious to learn why.

constructive 31 Dec 2023 9:05 UTC
1 point
0 ∶ 0
on: EU policymakers reach an agreement on the AI Act
At the very least, in my view, the picture has changed in an EU-favoring direction in the last year (despite lots of progress in US AI policy), and this should prompt a re-evaluation of the conventional wisdom (in my understanding) that the US has enough leverage over AI development such that policy careers in DC are more impactful even for Europeans.
Interesting! I don’t quite understand what updated you. To me, it looks like, given the EU AI Act is mostly determined at this stage, there is less leverage in the EU, not more. Meanwhile, the approach the US takes to AI regulation still remains uncertain, indicating many more opportunities for impact.

constructive 22 Aug 2023 12:08 UTC
1 point
0 ∶ 0
on: Collection of work on ‘Should you should focus on the EU if you’re interested in AI governance for longtermist/x-risk reasons?’
Thanks for the collection! (Note there is a typo in the title “Should you should focus on the EU if you’re interested in AI governance for longtermist/x-risk reasons?’”)

constructive 18 Jun 2023 9:49 UTC
2 points
0 ∶ 0
in reply to: pcihon’s comment on: A summary of current work in AI governance
Thanks for contributing these examples! Added a link to your comment in the main text.

constructive 26 Apr 2023 9:22 UTC
2 points
0 ∶ 0
on: [Cause Exploration Prizes] NOT Getting Absolutely Hosed by a Solar Flare
Thanks for the hint. Skimming this, it sounds somewhat exaggerated. I’d like to see a more rigorous investigation. (I.e., how strong can flares get, which equipment would be damaged.) This article suggests flares are much less harmful (only read first few paragraphs)

constructive 28 Mar 2023 7:49 UTC
5 points
1 ∶ 0
in reply to: MaxRa’s comment on: The Prospect of an AI Winter
(Uncertain) My guess would be that a global conflict would increase AI investment considerably, as (I think) R&D typically increases in war times. And AI may turn out to be particularly strategically relevant.

constructive 24 Mar 2023 19:14 UTC
3 points
0 ∶ 0
in reply to: sapphire’s comment on: Counterproductive Altruism: The Other Heavy Tail
Though you need to consider the counterfactual where the talent currently at OAI, DM, and Anthropic all work at Google or Meta and have way less of a safety culture.

constructive 25 Feb 2023 21:38 UTC
2 points
0 ∶ 0
in reply to: Felix Wolf’s comment on: Open Thread: January — March 2023
I think a central idea here is that superintelligence could innovate and thus find more energy-efficient means of running itself. We already see a trend of language models with the same capabilities getting more energy efficient over time through algorithmic improvement and better parameters/data ratios. So even if the first Superintelligence requires a lot of energy, the systems developed in the period after it will probably need much less.

constructive 18 Jan 2023 7:26 UTC
2 points
2 ∶ 0
in reply to: Patrick Gruban’s comment on: EA Germany’s Strategy for 2023
That sounds very sensible

constructive 17 Jan 2023 20:24 UTC
7 points
3 ∶ 0
in reply to: Patrick Gruban’s comment on: EA Germany’s Strategy for 2023
Weakly held opinion that you could be investing too much into this progress. I’d expect to hit diminishing returns after ~50-100 hours (though have no expertise whatsoever)

constructive 17 Jan 2023 20:22 UTC
2 points
1 ∶ 0
in reply to: Sarah Tegeler’s comment on: EA Germany’s Strategy for 2023
Sounds good! Thanks for the reply.
I think I would additionally find it helpful to get some insights into what you are prioritizing to give feedback on project plans (or maybe your future post on impact metrics will include that?) - but I know communication takes a lot of effort, so it may be easier to just receive that feedback as you’re rolling out the projects. Looking forward to your next post

constructive 17 Jan 2023 20:16 UTC
1 point
0 ∶ 0
in reply to: She's done it ’s comment on: Some intuitions about fellowship programs
Thanks!
I think even for community building, there is a rather broad category, though most of the ones I know of are online, so they probably don’t apply here. It would be helpful to have some examples.

constructive 17 Jan 2023 13:58 UTC
15 points
7 ∶ 0
on: EA Germany’s Strategy for 2023
Thanks for writing this!

Your high-level strategy sounds great but is also fairly generic, so hard to criticize. One theme throughout seems to be an emphasis on creating solid structures and making sure everything is legally backed. I think this is important-especially after the FTX crisis-though I sense a risk of an overemphasis on security. EA Germany is yet a small community and has so much room to grow! I think we need a large number of capable people to solve some serious problems fairly soon. EA Germany should thus emphasize ambition and growth, primarily through targeted outreach and connecting the community better internationally.

I was hoping to learn more about the concrete projects that you are working on/ about to start. Here are some ideas that I see as priorities (Most are already happening to some degree):
- Further growing Berlin as a hub, with the goal of making it attractive for people working full-time on EA causes. (E.g., organizing speaker events, creating housing for temporary stays, attracting talent)
- Connecting the German community better internationally: Inviting leading EAs to give talks in major German cities and encouraging Germans to attend international events. (I think one reason German EAs seem to be less ambitious than EAs from other countries is that they aren’t connected well to senior EAs and thus receive less mentorship and worse access to opportunities)
- Targeted outreach, particularly at national student fellowships such as Studienstiftung
- Creating pipelines for talent, e.g., by running AI alignment camps or local iterations of the AGISF
Thanks for all the work you are already doing! It’s always easy to suggest a lot of projects without having to start them myself :D
Looking forward to your strategy playing out.

constructive 17 Jan 2023 12:13 UTC
3 points
0 ∶ 0
on: Some intuitions about fellowship programs
Just for context—which programs are you referring to and have organized/ participated in? I see the term “fellowship” used for a range of programs

constructive 8 Dec 2022 16:59 UTC
1 point
0 ∶ 0
in reply to: constructive’s comment on: Have your timelines changed as a result of ChatGPT?
I think my skepticism comes from the fact that you could already easily cheat on my biology exams with internet access (and people did when they were online during covid). This isn’t much of a problem with pen & paper exams, but it seems that online degrees will really struggle now.

constructive 8 Dec 2022 16:57 UTC
1 point
0 ∶ 0
in reply to: Geoffrey Miller’s comment on: Have your timelines changed as a result of ChatGPT?
Thank you. I think for the first two I still don’t see a big jump from just internet access. But I agree that cheating got a lot easier. The third example seems particularly like outsourcing an important skill.

constructive 5 Dec 2022 19:20 UTC
5 points
3 ∶ 5
in reply to: Geoffrey Miller’s comment on: Have your timelines changed as a result of ChatGPT?
This sounds a bit exaggerated. I haven’t engaged much with ChatGPT, but it seems like the same arguments could be made for students having access to the internet.
Can you define what kind of writing assignments you are referring to?

constructive 15 Nov 2022 13:49 UTC
1 point
0 ∶ 0
in reply to: constructive’s comment on: Who wants to be hired? (May-September 2022)
[redacted]

constructive 11 Oct 2022 6:53 UTC
1 point
0 ∶ 0
on: Longtermists should take climate change very seriously
I think you’re basically right in your points, but they are not enough to say that climate change is nearly as bad as biorisk or AI misalignment. You may get close to nuclear risk, but I’m skeptical of that as well. My main point is that extinction from climate is much more speculative than from the other causes.

Reasons:
1. There is some risk of a runaway climate change. However, this risk seems small according to GWWC’s article + it would be overconfident to say that humanity can’t protect itself against it with future technology. There is also much more time left until we get to > 5° of warming than until the risk of engineered pathogens and powerful AI rises quickly.
2. Climate change will be very destabilizing. However, it’s very hard to predict the long-term consequences of this, so if you’re motivated by a longtermist framework, you should focus on tackling the more plausible risks of engineered pathogens and misaligned AI more directly. One caveat here is the perspective of cascading risks, which EA is not taking very seriously at the moment.
3. The impacts on life quality are not convincing from a longtermist standpoint as I expect them to last much less than 1000 years whereas humanity and its descendants could live for billions of years. I also expect only a tiny fraction of future sentiences to live on earth.
Another thought I often miss in debates on x-risk from climate change is that humans would likely intervene in climate at some stage if it’s a serious threat to our economies and even lives. I haven’t seen anyone make this point before, but please point me to sources.
If you are still new to EA, you may understand the current position better as you learn more about the pressingness of biorisk and especially AI risk. That said, there is probably room for some funding for climate change from a longtermist perspective, and given the uncertainty surrounding cascading risks, I’d be happy to see a small fraction of longtermist resources directed to this problem.

constructive 18 Aug 2022 20:50 UTC
2 points
0 ∶ 0
in reply to: Sjlver’s comment on: How to find *reliable* ways to improve the future?
Re: bioweapons convention: Good point, so maybe not as straightforward as I described.

Re: predicting AI: You can always not publish the research you are doing or only inform safety-focused institutions about it. I agree that there are some possible downsides to knowing more precisely when AI will be developed, but there seem to be much worse downsides to not knowing when AI will be developed (mainly that nobody is preparing for it policy- and coordination-wise)
I think the biggest risk is getting governments too excited about AI. So I’m actually not super confident that any work on this is 10x more likely to be positive.
Re: policy & alignment: I’m very confident, that there is some form of alignment work that is not speeding up capabilities, especially the more abstract one. Though I agree on interpretability. On policy, I would also be surprised if every avenue of governance was as risky as you describe. Especially laying out big picture strategies and monitoring AI development seem pretty low-risk.
Overall, I think you have done a good job scrutinizing my claims and I’m much less confident now. Still, I’d be really surprised if every type of longtermist work was as risky as your examples—especially for someone as safety-conscious as you are. (Actually, one very positive thing might be criticizing different approaches and showing their downsides)