RE: "i did get selected for the AI Safety Fellowship at Anthropic"
I'm helping out with the Anthropic Fellows Program and I think this is not the truth unless your name is not Vivek Singh as per your profile.
Do feel free to DM if I've misunderstood something!
writing this as clearly and as honestly as I can.
first, thank you so much for the love and support. and I'm sorry for the confusion I caused.
i advanced to the next round of MATS for the empirical track but i did not receive a final offer letter. that was my mistake in how I communicated it and i really own that and i am sorry for it. once I realized it myself I removed every mention of MATS from my profile because I wanted to try to be honest. I do have confrontation issues. I get that.
what I can confirm is that i did get selected for the AI Safety Fellowship at Anthropic. in a month or so ill share my experience and my work with you guys and i did had an internship at Starbucks. ill post the offer letter and the entire kit in a picture maybe and i also did get selected for TIFF and it was send out by an official tiff id
if you need proof, DM me and I'm happy to screenshare the emails or any credentials i just don't think it should be public and if i trust you shouldnt post it publicly too
i also felt dishonest about writing the "how to be good at research" article with AI assistance even if it was only for style, grammar, or polish. so many people have no problem with that. but I felt guilty. I had put in all the reading, all the references, tried to connect everything but using AI afterwards didn't feel right to me. I should have kept it in my own voice. I'm sorry for that too.
I won't be engaging in any further discussion. the hate has spread far and some people even doubted my panic attacks those are real, and have been since childhood.
I've never faced this kind of hatred before. but even through this, I want to thank @anpaure for making me realize something important which was when you accidentally do something wrong, clarify it openly. be more upfront. own it. although his post ruined everything but its totally fine.
I'll be focusing on my research. when I come back, I'll come back with solid work. I don't know exactly what's next I may step away for a while or dont continue with anything or just run away from everything i really have no idea i dont wanna put anything emotional because people in the comments will say i am a crybaby or i deserve it
but thank you for everything.
We are starting a new, nonprofit alignment organization, ⊢ Sequent Research, bringing together researchers previously on UK AISI’s Alignment Team, Timaeus, and elsewhere to research how to align superintelligence. We are hiring! 🧵
@YafahEdelman Hey! I don't have twitter premium so I've put a few Qs in a doc. Feel free to answer as many or few as interest you! Feel free to request edit access.
https://t.co/dzQblfb57q
@mishapathy Nah jk, in all seriousness I did consider adding caveats around not being conflict-averse, but thought the doc had too many footnotes already
Hey friends/strangers! I'll be moving to the Bay later this year.
Now that I expect to be less nomadic, I'll be looking for a long-term partner. I've always appreciated how earnest folks in this community are with date-me docs, so I thought I'd share mine too (link below).
@shaggysurvives People who have tens of thousands of followers actually have moral obligations to consider what they say, and at a certain point you become a public figure
Many people say things as bait and get what they deserve for it Nick decker is off the rails and suggests absurd things
An exciting milestone for AI in science: Our C2S-Scale 27B foundation model, built with @Yale and based on Gemma, generated a novel hypothesis about cancer cellular behavior, which scientists experimentally validated in living cells.
With more preclinical and clinical tests, this discovery may reveal a promising new pathway for developing therapies to fight cancer.
@AaronBergman18@absurdlymax If you think there's a discontinuity (any suffering involving fire, fear of death, psychosis, loss of autonomy), then you need categorical ES definitions that don't capture scenarios people would trivially make tradeoffs on for high wellbeing or BNES
@AaronBergman18@absurdlymax If you think "having the sensation of being burned alive maintained for a week" would cross the ES threshold, you could decrease the duration of time, the amount of nerve endings exposed to the burn, how often they fire, etc.
@tszzl@ESYudkowsky@zacharynado Any reason you're only considering PEPFAR rather than all USAID/global health funding?
Some estimates:
https://t.co/cJankyrpRF
https://t.co/nT7FeTpBHV
https://t.co/GAw1lFg42B
Even if these estimates are 5x off it's still in the range of hundreds of thousands of deaths