Not attending CS degree isn't a sacrifice. Self-learning is very sufficient, the only thing I need from school is research mentorship.
It's a bit complicated to explain, but doing these two in parallel does NOT reduce productivity at all. If you're interested at why I could explain in the future
The strategy of doing in parallel is: To attend dental school, but spending much less time on dental school than other students, getting like 2.5 GPA, just aim to pass and graduate in dental school. (In Taiwan, it's really feasible. Our dental school training is NOT rigid. In fact every time I went to dentist I always think maybe this dentist doesn't work hard in school)
Imagine: All of the students in dental would know there's a weird guy called Jack that is studying AI every day and never studies dental school seriously
Once you graduate in dental school, you can get the dental license at any time(even 10 years later) if you passed the exam. Therefore, for me after I graduated from dental school, if I later found myself better at EtG and not direct work, then I can self study dentistry again for years, pass the licence exam and start practicing.
You may ask: Why don't you study CS first, if 10 years later I found out myself isn't good enough in direct working, then I apply to dental school? Because I expect dental school would be much harder to apply in the future(which could be nearly impossible to apply)
I'm not saying I'm def nitely going to do this( the probability is around 35%), as I said with you in private messages, I'll test my personal fit for 2 years first. If I found myself bad at CS, then research mentorship wouldn't matter at all Because I won't do CS research in parallel. But if after 2 years I found myself seem to be good at direct work, but not confident then it seems it's the best plan)
Hello Geoffrey: Thanks for your reply, and I'm also asking the professors in Taiwan I know on whether they would be open to collabortae with students from other colleges. You're right that this question is somewhat localized. EA people in Taiwan that's familiar with doing AI risks reserach would be the best to answer this question, however there are very few(<3 people) in Taiwan. Therefore I think it'd still be beneficial to see how western EA thinks about this question
I have a question related to this post: https://forum.effectivealtruism.org/posts/wX6maSvjHMyCHAoui/is-pursuing-high-gpa-aligned-with-training-ai-risks-research
For people who disagreed him: Just disagree, don't downvote. Every reply takes time and should be respected. I think downvote should only be used to posts/replies that are obvi usly not serious
Hello Ariel: Thanks very much for your response, it's very helpful.
However, for my situation it's a bit different: If I double major in dentistry and CS(I'll probably do them in different colleges, the below is just discussing on the CS part). I could choose a top college in Taiwan. However, in the top college I'd have to invest like much more time in order to get a top 10-20% GPA(like 3.8). In applying master in USA, it seems GPA matters much more than your college name(especially because they don't know how big the difference of students intelligence is between the top-teir and mid-teir school in Taiwan.) Therefore, there's a case for choosing mid-teir school in order to spend less time and maintain high GPA.
So the choice may be like: 1.spend 1000 hours extra on chasing GPA, and get a 3.2 GPA in a top-tier school 2.spend 1000 hours extra on chasing GPA, get a 3.8 GPA in a mid-tier school 3.spend 2000 hours extra on chasing GPA, get a 3.8 GPA in a top-tier school. It seems obvious that 1 is the worst for applying master degree in USA, so the choice is between 2 and 3, whether it's worth investing like 1000 hours to improve your "college school name" on the resume. What do you think?
(And of course, there are still other advantages of choosing a top-tier school, like better peers and research environment.)
You're right that E2G depends on cause area. However, I doubt the saying "AI safety isn't funding constrainted", as I see the funding bar is still really high.(and some aspiring AI safety researchers can't get an EA job). It's true that some org said they don't need more money, but they could be too stingy on funding researchers. In the AI s-risks(suffering risks) field, it's probably more funding constrainted than AI x-risks field. Because, if there are only 1 or 2 major donors in a field, there maybe blind spots of the donor and some great projects aren't funded. As I know, there are very few indidvidual donors donating > $50000 annually now. In this case, small donors could fill the gap. Imagine if there's a PR90 researcher unfunded, if you donate to him, your impact would be better than 90% of people in EA. I have a FB post on EA career discussion group to explain deeper of this idea:
People intersted in this topic can also read this: https://forum.effectivealtruism.org/posts/tMv3KWqcx5oFq27gN/the-contribution-of-average-direct-worker-in-ai-risks-may-be
IMO the fact is crazier, it's "despite billions of funding, small donors still have a siginificant better marginal value of dontaing money than big donors"
Not attending CS degree isn't a sacrifice. Self-learning is very sufficient, the only thing I need from school is research mentorship.
It's a bit complicated to explain, but doing these two in parallel does NOT reduce productivity at all. If you're interested at why I could explain in the future
Hello Peter:
The strategy of doing in parallel is: To attend dental school, but spending much less time on dental school than other students, getting like 2.5 GPA, just aim to pass and graduate in dental school. (In Taiwan, it's really feasible. Our dental school training is NOT rigid. In fact every time I went to dentist I always think maybe this dentist doesn't work hard in school)
Imagine: All of the students in dental would know there's a weird guy called Jack that is studying AI every day and never studies dental school seriously
Once you graduate in dental school, you can get the dental license at any time(even 10 years later) if you passed the exam. Therefore, for me after I graduated from dental school, if I later found myself better at EtG and not direct work, then I can self study dentistry again for years, pass the licence exam and start practicing.
You may ask: Why don't you study CS first, if 10 years later I found out myself isn't good enough in direct working, then I apply to dental school? Because I expect dental school would be much harder to apply in the future(which could be nearly impossible to apply)
I'm not saying I'm def nitely going to do this( the probability is around 35%), as I said with you in private messages, I'll test my personal fit for 2 years first. If I found myself bad at CS, then research mentorship wouldn't matter at all Because I won't do CS research in parallel. But if after 2 years I found myself seem to be good at direct work, but not confident then it seems it's the best plan)
Hello Geoffrey: Thanks for your reply, and I'm also asking the professors in Taiwan I know on whether they would be open to collabortae with students from other colleges. You're right that this question is somewhat localized. EA people in Taiwan that's familiar with doing AI risks reserach would be the best to answer this question, however there are very few(<3 people) in Taiwan. Therefore I think it'd still be beneficial to see how western EA thinks about this question
The part about "EA reinventing the wheel" is quite insipring, thanks for providing me this opinion.
I have a question related to this post: https://forum.effectivealtruism.org/posts/wX6maSvjHMyCHAoui/is-pursuing-high-gpa-aligned-with-training-ai-risks-research
For people who disagreed him: Just disagree, don't downvote. Every reply takes time and should be respected. I think downvote should only be used to posts/replies that are obvi usly not serious
Hello Ariel: Thanks very much for your response, it's very helpful.
However, for my situation it's a bit different: If I double major in dentistry and CS(I'll probably do them in different colleges, the below is just discussing on the CS part). I could choose a top college in Taiwan. However, in the top college I'd have to invest like much more time in order to get a top 10-20% GPA(like 3.8). In applying master in USA, it seems GPA matters much more than your college name(especially because they don't know how big the difference of students intelligence is between the top-teir and mid-teir school in Taiwan.) Therefore, there's a case for choosing mid-teir school in order to spend less time and maintain high GPA.
So the choice may be like: 1.spend 1000 hours extra on chasing GPA, and get a 3.2 GPA in a top-tier school 2.spend 1000 hours extra on chasing GPA, get a 3.8 GPA in a mid-tier school 3.spend 2000 hours extra on chasing GPA, get a 3.8 GPA in a top-tier school. It seems obvious that 1 is the worst for applying master degree in USA, so the choice is between 2 and 3, whether it's worth investing like 1000 hours to improve your "college school name" on the resume. What do you think?
(And of course, there are still other advantages of choosing a top-tier school, like better peers and research environment.)
Hello Ariel
You're right that E2G depends on cause area. However, I doubt the saying "AI safety isn't funding constrainted", as I see the funding bar is still really high.(and some aspiring AI safety researchers can't get an EA job). It's true that some org said they don't need more money, but they could be too stingy on funding researchers. In the AI s-risks(suffering risks) field, it's probably more funding constrainted than AI x-risks field. Because, if there are only 1 or 2 major donors in a field, there maybe blind spots of the donor and some great projects aren't funded. As I know, there are very few indidvidual donors donating > $50000 annually now. In this case, small donors could fill the gap. Imagine if there's a PR90 researcher unfunded, if you donate to him, your impact would be better than 90% of people in EA. I have a FB post on EA career discussion group to explain deeper of this idea:
https://www.facebook.com/share/p/18oL5QjUUq/
People intersted in this topic can also read this: https://forum.effectivealtruism.org/posts/tMv3KWqcx5oFq27gN/the-contribution-of-average-direct-worker-in-ai-risks-may-be
IMO the fact is crazier, it's "despite billions of funding, small donors still have a siginificant better marginal value of dontaing money than big donors"