MathiasKB🔸

Karma: 6,508

MathiasKB🔸Mar 31, 2024, 9:23 AM
4 points
0 ∶ 0
in reply to: Geoffrey Miller’s comment on: Geoffrey Miller’s Quick takes
I haven’t seen the series, but am currently halfway through the second book.
I think it really depends on the person. The person I imagine would watch three-body problem, get hooked, and subsequently ponder about how it relates to the real world, seems like they also would get hooked by just getting sent a good lesswrong post?
But sure, if someone mentioned to me they watched and liked the series and they don’t know about EA already, I think it could be a great way to start a conversation about EA and Longtermism.

MathiasKB🔸Mar 29, 2024, 8:44 AM
13 points
2 ∶ 0
on: [Linkpost] Leif Wenar’s The Deaths of Effective Altruism
Relevant to the discussion is a recently released book by Dirk-Jan Koch who was Chief Science Officer in the Dutch Foreign Ministry (which houses their development efforts). The book explores the second order effects of aid and their implications for an effective development assistance: Foreign Aid And Its Unintended Consequences.
In some ways, the arguments of needing to focus more on second-order effects are similar to the famous ‘growth and the case against randomista development’ forum post.
The west didn’t become wealthy through marginal health interventions, why should we expect this for Sierra Leone or Bangladesh?
Second-order effects are important and should be taken into as much consideration as the first-order effects. But arguing that second-order effects are more difficult to predict, and we therefore shouldn’t do anything falls prey to the Copenhagen Interpretation of Ethics.

MathiasKB🔸Mar 29, 2024, 8:21 AM
5 points
1 ∶ 0
in reply to: Henry Howard🔸’s comment on: [Linkpost] Leif Wenar’s The Deaths of Effective Altruism
just fyi Dean Karlan doesn’t run USAID, he’s Chief Economist. Samatha Power is the (chief) administrator of USAID.

MathiasKB🔸Mar 13, 2024, 11:16 AM
13 points
3 ∶ 0
on: The RCT Agenda—Bryan Caplan
I think Bryan Caplan is directionally correct, but his argumentation in this post is incredibly sloppy.
A marxist communist could make the exact same complaint as Bryan Caplan, but with the signs flipped. Why do all these economists focus on RCTs for educational interventions, and never once consider the best educational intervention is to rise up in violent revolution and overthrow our capitalist oppressors?
I don’t recall any of the RCT papers I’ve read being particularly heavy on normative claims. Usually they’ll just say:
“this intervention had a measurable effect on X, so policy makers interested in improving X should consider it part of the tool kit”
or
“this intervention didn’t have an effect on X, so policy makers interested in improving X should not do this”
Which seems completely reasonable to me. They aren’t quietly rejecting the question, they largely are just not engaging with normative questions of what policy makers ought to do. RCTs are a way of taking out ideology and focus on strictly empirical questions.

MathiasKB🔸Mar 11, 2024, 12:03 PM
6 points
1 ∶ 0
on: I am unable to get any AI safety related fellowships or internships.
Consider joining hackathons such as the ones organized by Apart Research. Anyone can join and get to work on problems directly related to AI Safety.
If you do a good project, you can put that on your resume and have something to speak about at your next interview.

MathiasKB🔸Mar 7, 2024, 10:30 AM
12 points
2 ∶ 0
on: Why are you reluctant to write on the EA Forum?
I think there’s at least two categories:
1. The beginner who is scared of ridicule.
2. The senior, who don’t have time to write to the forum standard without risking reputation.
I’m more interested in what we can do to encourage the latter group. My impression is that many senior people are reluctant to post, as they don’t have time to write something sufficiently well-argued and respond to the comments.
Instead many good discussions take place in signal groups, google docs and email threads. In a perfect world, these discussions would be in the forum. The issue right now is that if those conversations would take place on the forum, too many people chime in with long, eloquently written, but wrong arguments that the subject matter experts now have to spend additional time shutting down or look like they are dodging a hard question.
Additionally lowering their bar for public engagement, puts them at risk of attack. There are people reading this forum just to find ammunition for hit pieces. A poorly worded comment from the leader of an organization will be used against them.

MathiasKB🔸Mar 5, 2024, 2:44 PM
13 points
1 ∶ 1
on: MathiasKirkBonde’s Shortform
Does Claude-3 push capabilities?
I think it can be a fun exercise is to just interpret CEOs statements literally and see what they imply.
If Dario Amodei claims they don’t want to push capabilities, I think an interesting question to ask is in what sense releasing the world’s best LLM isn’t pushing capabilities.
One option that seems possible to me, could be that they no longer consider releasing improved LLMs to meaningfully push the frontier. If Claude-3 spurs OpenAI to push a quicker release of GPT-4.5, this would not be an issue as releasing ever more refined LLMs doesn’t meaningfully increase any specific risk.
In that case they would still be true to their word, Anthropic just no longer considers LLMs on their own to be the frontier of capabilities. This seems reasonable to me?

MathiasKB🔸Mar 3, 2024, 7:51 PM
4 points
1 ∶ 0
on: How EA can be better at communications
I thought the video was excellent, and the highlights of your article were the concrete ideas and examples of good communication.
More concrete ideas please! I don’t think anyone will disagree that EA hasn’t been the best at branding itself, but in my experience it’s easier said than done!

MathiasKB🔸Mar 1, 2024, 9:52 AM
10 points
1 ∶ 0
on: Creative video ads significantly increase GWWC’s social media engagement and web traffic to pledge page
Really cool experiment!
Was it possible to track to what extent the more engaging ads drove conversions? (donations made, pledges taken, etc.)
My hypothesis would be the more engaging ads get more people onto the website, but those people will be much less likely to follow through (and especially with significant amounts), than for example a very targeted and nerdy ad aimed at wealthy tech workers.

MathiasKB🔸Feb 20, 2024, 4:13 PM
3 points
2 ∶ 0
on: Ten Commandments for Aspiring Superforecasters
I think this leaves out what is perhaps the most important step in making a quality forecast, which is to consider the baserates!

MathiasKB🔸Feb 17, 2024, 4:11 PM
2 points
0 ∶ 0
on: MathiasKirkBonde’s Shortform
Signal boosting my twitter poll, which I am very curious to have answered:
https://twitter.com/BondeKirk/status/1758884801954582990

Basically the question I’m trying to get at is whether having hands-on experience training LLMs (proxy for technical expertise) makes you more or less likely to take existential risks from AI seriously.

MathiasKB🔸Feb 14, 2024, 9:51 AM
19 points
12 ∶ 0
in reply to: JoshuaBlake’s comment on: FTX expects to return all customer money; clawbacks may go away
and even if they were solvent at the time, that does not mean they were not fraudulent.
If I took all my customers money, which I had promised to safekeep, and went to the nearest casino and put it all on red, even if I won it would still be fraud.

MathiasKB🔸Feb 9, 2024, 9:29 AM
9 points
0 ∶ 0
on: Celebrating Benjamin Lay (died on this day 265 years ago)

MathiasKB🔸Dec 20, 2023, 7:14 PM
12 points
2 ∶ 1
on: It is called Effective Altruism, not Altruistic Effectiveness
In conclusion, I think that rather than being overly focused on finding the most effective means of doing good, we should also be concerned with becoming more altruistic, caring and compassionate.
I strongly agree with the last half of this sentence. A rocket engine is only valuable insofar as it is pointed in the right direction. Similarly to how it makes sense to practice using spreadsheets to systematize ones decision-making, I think it make sense to think about ways to become more compassionate and kind.

MathiasKB🔸Dec 18, 2023, 9:40 AM
5 points
7 ∶ 2
on: OpenAI, DeepMind, Anthropic, etc. should shut down.
We do not know how to make a PAI which does not kill literally everyone.
We don’t know how to make a PAI that does kill literally everyone either. What would the world have to look like for you to be pro more AI research and development?

MathiasKB🔸Dec 16, 2023, 11:21 AM
8 points
1 ∶ 0
in reply to: SiebeRozendal’s comment on: SiebeRozendal’s Shortform
Just did it, still works. You can donate to what looks like any registered US charity, so plenty of highly effective options whether you care about poverty or animal welfare.

MathiasKB🔸Dec 12, 2023, 4:35 PM
3 points
0 ∶ 0
in reply to: Nathan Young’s comment on: Nathan Young’s Shortform
There’s a few I know of:
- For the new R21 vaccine, WHO is currently conducting prequalification of the production facilities. As far as I understand, African governments have to wait for prequalification to finish for before they can apply for subsidized procurement and rollout through UNICEF and GAVI.
- For both RTS,S and R21, there are some logistical difficulties due to the vaccines’ 4 dose schedule (First three 1 month apart—doesn’t fit all too well into existing vaccination schedules) cold-chain requirements, and timing peak immunity with the seasonality of malaria.
- Lastly since there already exists cost-effective counter-measures, it’s unclear how to balance new vaccine efforts against existing measures.

MathiasKB🔸Nov 30, 2023, 8:21 PM
9 points
0 ∶ 0
in reply to: Benny Smith’s comment on: ludwigbald’s Shortform
Ah, today I learned! thanks for correcting that. For what it’s worth I was vegan for two years, and have been vegetarian for 6.
Do you happen to know about the bioavailability claims of animal versus plant protein?

MathiasKB🔸Nov 30, 2023, 6:23 PM
0 points
0 ∶ 3
in reply to: Benny Smith’s comment on: ludwigbald’s Shortform
They literally don’t. Animal proteins contain every essential amino acid, whereas any plant protein will only have a subset.

MathiasKB🔸Nov 30, 2023, 12:57 PM
3 points
0 ∶ 1
in reply to: ludwigbald’s comment on: ludwigbald’s Shortform
I’m quite excited about cricket protein! Nutritionally it’s superior to vegan protein supplements, especially for people who are otherwise vegan and won’t get animal protein.
My intuition is that it very much comes down to whether one views an undisturbed cricket life as net-positive or negative. A cricket farm breeds millions of crickets in a 6 week cycle where the crickets are frozen to death ~~not long before they naturally would die of old age~~.
Rethink Priorities recently incubated the insect institute who I think are exploring insect sentience. They’re more qualified to speak on this than I am.
EDIT: turns out I don’t know shit about crickets or nutrition. Rethink has a cool report on insect farming, also points out my claim on their death being soon before natural old age is likely wrong. https://forum.effectivealtruism.org/posts/ruFmR5oBgqLgTcp2b/insects-raised-for-food-and-feed-global-scale-practices-and#Cricket_farming_practices_and_conditions