Daniel Sosebee (@dnsosebee) header
Daniel Sosebee (@dnsosebee) profile picture

trying to live up to my values

San Francisco
Joined June 2021
Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)Photo by Daniel Sosebee (@dnsosebee)
Daniel Sosebee retweeted
I've had some great moments with aloud, my voice-driven meditation app lately - and now it's in limited beta for Android and iOS. I need 12 testers to sign up for Android specifically so I can go live as a Real App. Extra free credits for mobile testers! aloud.rest/mobile-beta/
7
1
9
508
If you want better parties, you have to make better parties. Enter: Maximum Republicans. New York needs a competitive political ecology, and two parties that compete to offer a strong, positive vision of governance. Here's my part of that grand project: maximumnewyork.com/p/maximum…
34
23
42
457
97,145
Foundations of the Liberal Arts, Cohort 3, applications will open this week. It will start in about three weeks, and will be Sunday evening to give you a refreshing start to the week.
1
2
11
962
Daniel Sosebee retweeted
You know, @OpenAI, I heard a rumor @AnthropicAI solved alignment and they're publishing in two weeks.
53
211
20
4,188
93,726
Daniel Sosebee retweeted
try this app i made that makes your phone less addictive!
7
5
2
48
5,791
Daniel Sosebee retweeted
Seeing reactions like "Frequent scanning is bad, over-testing leads to unnecessary procedures that cost thousands of dollars." This take is catastrophically wrong. More data is unequivocally good. We should adjust the threshold for follow-on procedures, not make ourselves blind.
Announcing a new division of Midjourney called "Midjourney Medical"
72
45
24
1,020
157,916
Daniel Sosebee retweeted
i made a website where I'm detailing experiments and what I've learned so far. ideas, feedback, etc appreciated! aphonethatcares.com
2
3
39
1,316
Daniel Sosebee retweeted
where would i look to find a room for a month in the bay? anyone have connections?
2
2
16
1,745
Daniel Sosebee retweeted
To add more context, as this post is getting a lot of views: Leading up to this 6.6 mile run I had already resumed running and climbing regularly. My runs until this point has been in the 2-3 miles range, and always included some walking (as did the run above). So: this run was a big achievement, as was the week of activity that followed, but - none of it defies expectations of a normal optimistic progression back towards physical health (especially considering that I was a runner and a climber before the illness). The point is: What is special here is that I proved to myself that a particular case of fatigue I was experiencing turned out to be inconsequential, and I proved that I can proceed at a normal and even optimistic rate of recovery, which had previously seemed unlikely due to all the symptoms I would feel around exercising. I did not magically get back to normal physical health, I had already put in a lot of gradual work over the preceding months (walking, daily yoga, gradual introduction to running and climbing). For all of the relevant context, please read the post itself, which goes through my recent health history, especially as it pertains to exercise.
1
1
3
432
Daniel Sosebee retweeted
This feels very similar to the best ways that I know to recover from the nervous system sensitization that I get. I don't have long COVID but my nervous system perpetually contracts its window of tolerance until I can't handle a lot of normal things.
4
1
11
987
Daniel Sosebee retweeted
file under "Rhymes with John Sarno" (Under some conditions, some people recover by just "pretending they're not sick." How many? What conditions? Idk but it's fascinating)
Replying to @dnsosebee
I wrote up all the details (link in next)
1
7
169
Daniel Sosebee retweeted
Replying to @0xSigil
Bro, this is wrong. Lengthening the feedback distance between humans and AIs is not a good thing for the world. Today, it means you're generating slop instead of solving useful problems for people. It's not even well-optimized for helping people have fun. Once AI becomes powerful enough to be truly dangerous, it's maximizing the risk of an irreversible anti-human outcome that even you will deeply regret. The point of ethereum is to set *us* free, not to create something else that goes off and does some stuff freely while our own situation is unchanged or worsened. (And, as others have pointed out, the models are run by openai and anthropic, so the thing is not even "self-sovereign"; you're actually perpetuating the mentality that centralized trust assumptions can be put in a corner and ignored, the very mentality that ethereum is at war with) The exponential will happen regardless of what any of us do, that's precisely why this era's primary task is NOT to make the exponential happen even faster, but rather to choose its direction, and avoid collapse into undesirable attractors.
399
295
204
3,503
569,295
BOOK LAUNCH 🎉! My new book of drawings is now available for purchase. It includes over 250 of the 680+ drawings I've posted here on twitter over the past 2 years. Link is in the replies, plus some example drawings and some kind words that people have shared about my drawings.
28
31
18
311
53,756
What I've done for government with Maximum New York, I will also do for the liberal arts via The Algernon Project. New fall class: Foundations of the Liberal Arts
3
2
2
33
2,917
Join us on August 13 for Localhost! 🦠 @computerender will present Mote, an interactive ecosystem simulation with hundreds of thousands of organisms. His custom GPU physics engine models many simple behaviors at a massive scale, producing fascinating emergent phenomena. RSVP⬇️
24
169
30
1,398
107,838
Daniel Sosebee retweeted
An update on Fractal Tech Bootcamp, and a request for help 🙏🏼 I started an AI engineering accelerator (a “bootcamp”) about 1 year ago, because my friends and I could not find enough good AI talent to hire, and I’m obsessed with educational reform. The bootcamp was an experiment- an attempt to see how much better adult education can be when we apply principles from my education research into Montessori, Xerox PARC, ARPA, The Manhattan Project, Black Mountain College, YCombinator, The Royal Society, The Jesuits, and other great cultures. It might sound grandiose, but what if I treated culture and education as an engineering project? What if I leveraged everything I knew about tech, culture, and education to help students learn effectively? We've run 3 cohorts so far, and the results are very good. Even in a tough tech market, we have a 100% placement rate into SWE jobs. One of my students raised $500k for his startup. And the thing is, this is barely a prototype- I’ve only built out a fraction of the educational program I intend to build. I'm just getting started. Montessori describes a process called “normalization” by which the student adapts to the culture of focused work in the classroom. After a child has normalized, they work just for the sake of it, because they want to, because they enjoy it, because they love the work of learning, because everyone around them loves the work of learning. We have built a space and a culture where everyone loves to work on software engineering, all the time. At Fractal Tech, students are immersed in this rigorous, high-achieving environment where shipping 3-5 PRs per day is a baseline expectation of the culture, because we enjoy being prolific. Students learn how to work on professional engineering teams before they graduate, because every student gets an internship with a local startup. Students learn what real code review looks like, because every instructor has led engineering teams, and we deliver brutally honest (and invaluable) feedback on code quality. Also, we built an entire tech coworking space to create a feeling of immersion. If you want to teach someone French, don't give them a dictionary, just drop them in France. Students work side-by-side with over 50 other professional founders and engineers, who naturally mentor the students and show them what real work looks like. Students work in-person, they make friends, they network casually, just by being in the space. I’m proud of what we've built so far, and now my goal is to share it with the world. I’m reaching out to Twitter for referrals because you know what I’m about, and I could use your help finding ideal students. Maybe you, reading this, are my ideal student :0 If you have a friend who could use our program, and our network, to level up as a software engineer, please refer them to me. Our ideal student knows how to code, but wants help becoming great: Think recently unemployed devs ready to bounce back, self-taught devs or CS students looking to break into the industry, coders and PMs stuck in boring jobs, or entrepreneurs who need technical skills & a network to make their startup happen. They will be in good hands. Sincerely, Andrew P.S. website in next tweet to appease the Twitter gods
7
39
12
256
61,920
Two of my favorite writers on the internet @patio11 & @danielgolliher just dropped a podcast episode together.
3
5
2
27
6,701
Daniel Sosebee retweeted
a soft contemplation
Replying to @dnsosebee
Sometimes, I sit down with a tune I'm not sure about, and it just develops into something wonderful, out of nowhere. Today was one of those days. Working on the 11th and probably final song in the project. Take a listen 🎹🎶: samply.app/p/Ag6Amerv6jOwVqN…
1
1
100
Replying to @DaveShapi
Thanks for engaging. Some responses (mine only, not anyone else on AI 2027 team): >>> "1. They provide no evidence that they guessed right in the past". In the "Who Are We?" box after the second paragraph, where it says "Daniel Kokotajlo is a former OpenAI researcher whose previous AI predictions have held up well", the words "previous AI predictions" is a link to the predictions involved, and "held up well" is a link to a third-party evaluation of them. >>> "2. No empirical data, theory, evidence, or scientific consensus". There's a tab marked "Research" in the upper right. It has 193 pages of data/theory/evidence that we use to back our scenario. >>> "3. They pull back at the end saying "We're not saying we're dead no matter what, only that we might be, and we want serious debate" okay sure." Our team members' predictions for the chance of AI killing humans range from 20% to 70%. We hope that we made this wide range and uncertainty clear in the document, including by providing different endings based on these different possibilities. >>> "4. The primary mechanism they propose is something that a lot of us have already discussed (myself included, which I dubbed TRC or Terminal Race Condition). Which, BTW, I first published a video about on June 13, 2023 - almost a full 2 years ago. So this is nothing new for us AI folks, but I'm sure they didn't cite me." Bostrom discusses this in his 2014 book (see for example box 13 on page 303), but doesn't claim to have originated it. This idea is too old and basic to need citation. >>> 5. "They make up plausible sounding, but totally fictional concepts like "neuralese recurrence and memory" Neuralese and recurrence are both existing concepts in machine learning with a previous literature (see eg nlp.cs.berkeley.edu/pubs/And… and en.wikipedia.org/wiki/Recurr…). The combination of them that we discuss is unusual, but researchers at Meta published about a preliminary version in 2024, see arxiv.org/pdf/2412.06769. We have an expandable box "Neuralese recurrence and memory" which explains further, lists existing work, and tries to justify our assumptions. Nobody has successfully implemented the exact architecture we talk about yet, but we're mentioning it as one of the technological advances that might happen by 2027 - by necessity, these future advances will be things that haven't already been implemented. >>> 6. "In all of their thought experiments, they never even acknowledge diminishing returns or negative feedback loops." In the takeoff supplement, CTRL+F for "Over time, they will run into diminishing returns, and we aim to take this into account in our forecasts." >>> 7. "They do acknowledge that some oversight might be attempted, but hand-wave it away as inevitably doomed." In one of our two endings, oversight saves the world. If you haven't already, click the green "Slowdown" button at the bottom to read this one. >>> 8. "Interestingly, I actually think they are too conservative on their 'superhuman coders'. They say that's coming in 2027. I say it's coming later this year." I agree this is interesting. Please read our Full Definition of the "superhuman coder" phase at ai-2027.com/research/takeoff… . If you still think it's coming this year, you might want to take us up on our offer to bet people who disagree with us about specific milestones, see docs.google.com/document/d/1… . >>> "9. Overall, this paper reads like 'We've tried nothing and we're all out of ideas.'" We think we have lots of ideas, including some of the ones that we portray as saving the world in the second ending. We'll probably publish something with specific ideas for making things go better later this year.
25
51
8
1,040
52,377