Amanda Ngo
@manda_ngo
Followers
2K
Following
928
Media
25
Statuses
307
relearning how to do stuff with presence • REMEMBERING TO PLAY • https://t.co/4uIVpXaEfo Map of the healing space: https://t.co/QjFXmODu4e
SF
Joined November 2017
I wrote a thing, along with the amazing @joelbot3000 . I'm really proud of it. It's the culmination of a year's worth of wandering around the space of AI governance & wellbeing in awe and confusion.
We hear a lot about “beneficial AI” or AI “aligned with human values” — but what does this mean and how do we get there? @joelbot3000 and @manda_ngo write:
1
3
34
This makes me want to hug GPT-3 and believe in myself <3.
one of the best things to come out of @oughtinc hackathon today: @manda_ngo teaching gpt3 to reframe negative thoughts positively. cc @elicitorg
4
22
182
so excited to be part of the team that won the @foresightinst existential hope & transformative AI hackathon $10k grant! We shared our vision for the Flourishing Foundation, designed to direct AI towards what really matters: our well-being & connection w/ self, others, & planet.
3
4
54
1/ Launching 2 new lit review tasks in @elicitorg:. Instead of Googling for hours, give Elicit a topic + examples and get back questions, key terms, and papers you should know about in minutes. Try it out ->
2
11
50
I'm currently hiring for a special ops ML intern to do rapid prototyping work testing the limits of GPT-3 in a research / Q&A setting (e.g. can we use GPT-3 to generate potential confounds in a research study?). DM me if this sounds exciting.
E.g. what's the most advanced task you can get GPT-3 to perform? What's the least advanced task it can't? Large language models are so new that you don't need much context to learn something about them that nobody has ever discovered. Then write about it!.
1
13
47
New in @elicitorg: show just the key claims of a paper. It's fun to answer my questions by skimming a few sentences
1
6
34
Had an incredible time putting together the Wellbeing <> AI Build Weekend. Huge thank you to @fiftyyears for hosting, & @OpenAI & @AnthropicAI for credits & model access! Was inspiring to glimpse what the future of wellbeing & AI could look like
1
7
30
Very excited about this!. Having concrete tasks humans care about where LLMs won’t do what we want by default is a helpful step towards having an AI Alignment benchmark we can test diff solutions on.
We’re announcing the Inverse Scaling Prize: a $100k grand prize + $150k in additional prizes for finding an important task where larger language models do *worse*. Link to contest details: 🧵
0
1
29
We made it easier & faster to go from a research question to relevant questions & papers in @elicitorg. It’s starting to feel delightful. Try it out!
1
3
22
had a glorious convo with @nopranablem and got inspired and made this. i think it's going to be fun! come through if u want to try ideal parent figure. (oo oo, i am also really excited about this ipf + parts work approach I have been experimenting with,.
1
1
20
@aadithyacdas yes it is astounding how long it takes me to realize. like I'm like wow I'm kind of depressed what's going on and then I look back on my week and realize I haven't had a consistent sleep / wake up time in 6 days and I'm like ohhhh.
1
1
17
.@elicitorg now gives you the one sentence from each abstract that's most relevant to your question:
3
6
17
@Conaw @RoamResearch So glad you found this Conor!! Have been very happily using Roam and wanted to share that with people!.
2
2
15
@robertwiblin Paddy Power and PredictIt also predict fewer electoral college votes for Biden than 538:
1
3
15
this week's launch: top papers in @elicitorg are better! we now rank by citation count, recency, and whether the paper answers your question
0
5
14
I was extremely astounded how much sleep masks made a difference in an already pretty dark room. Wonder how much cognitive overhead & underlying stress it adds to filter out noise and light all the time.
underdiscussed potential benefit of remote work: employees become more productive as they move to less urban areas because less noise and light pollution disrupting sleep @manda_ngo.
2
1
13
We're particularly excited about a well-being certification: a public eval & rating of the effect of AI products on our wellbeing. Here's a janky demo we made by evaling Reddit threads about @youper_ai (~AI therapist) and @MyReplika (~AI gf) w/ the UK wellbeing dashboard metrics
1
0
12
I have been so lucky to get to experience Dennis' coaching a handful of times. I have been struck by his patience, his non-agenda'dness, and his humility to let whatever is there unfold. I felt such a sense of loving, patient companionship by my side as I went to all these.
“What is it you plan to do with your one wild and precious life?”. These last few years I’ve been guided by the living question of purpose. The answer continues to evolve (and shorten) as life unfolds — maybe I’m nearing the essence. How can I guide people toward freer.
1
2
11
So grateful to @foresightinst and @FLI_org for putting this together and cultivating such a wonderful group of people, and also @jelennal_, @isahampt , @MingzhuHe1 and Chiara Gerosa for being a fire team to work with!.
1
1
11
Hmu if you’re a designer and want to help us figure out how to make research & truth-finding feel fast, intuitive, delightful (it is really quite fun).
So @jungofthewon gave me the breakdown of today. If you're a designer thinking about the future. The opportunity here to design the composition tools for building a personal, more human AI assistant is huge. What an insanely fun opportunity.
1
1
10
apps for this close tomorrow!. I did this last year. It's hard to put in words how precious of an experience this was, but to attempt: ~11 incredible facilitators are there to hold space for and meet anything coming up in you with love, for a month, with profound safety and.
Had a lovely walk with the founder of Sleep Awake. I really admire how much thought they are putting into supporting attendees during and post event. Sleepawake Camp is focused on orienting 18 - 30 yr olds towards a sense of wholeness and authentic happiness through science,.
0
1
8