"Create clarity at every zoom level" is one succinct way to understand the value of Asana.
Pinned Tweet
Show this thread
Follow
Dustin Moskovitz
@moskov
CEO & co-founder . Cosmopolitan consequentialist . Previously, co-founder . Openly mischievous, but secretly good.
Dustin Moskovitz’s Tweets
Covid vaccines saved about 20 million lives.
Quote Tweet
Global impact of the first year of COVID-19 vaccination: a mathematical modelling study thelancet.com/journals/lanin
286
785
4,158
Show this thread
New gene variant discovered that seems to protect against familial Alzheimer's. Much to follow up on, indeed a whole new avenue to explore...! All thanks to a generous Colombian family: mapp.mgh.harvard.edu/projects/colbo
(We support & team's work via )
Quote Tweet
Important lesson from nature.
A man genetically destined to manifest familial Alzheimer's disease had a rare gene variant that may/likely accounts for the protection
nature.com/articles/s4159 @NatureMedicine by @ytquiroz and colleagues
1
24
For the good of humanity, I spent my morning* open sourcing the YudBot2000 in hopes that someone will automate it for me.
?
poe.com/YudBot2000
* the last 2 mins
4
1
5
Poe bots are a great no-code middle way for programming an LLM-based app. Effectively they just save a long pre-prompt for you that you can refine over time instead of retyping all your best practices each time you use a chatbot.
Quote Tweet
Please remember that our API is targeted at developers who know how to write at least a small amount of code. For those who don’t, we have a separate way to create a bot on Poe based on a simple prompt, which is accessible at poe.com/create_bot.
Show this thread
3
3
27
Huge win! The Supreme Court just upheld California’s Prop 12, the nation’s strongest farm animal welfare law.
This ruling spares millions of animals from crates and sets a powerful precedent that could help many more.
More details ⬇️
supremecourt.gov/opinions/22pdf
39
366
1,435
Show this thread
Announcing Helion’s first customer: Microsoft.
We expect to start producing electricity in the world’s first fusion power plant by 2028, dramatically shortening the timeline for commercially viable fusion energy.
Read more:
118
1,092
3,240
microsoft becomes helion's first customer, in the first commercial deal for fusion power:
271
1,086
5,999
Because it is *your* agent, running locally, it’s a data privacy win too. A wallet for unstructured info. Give up as little as you want to each vendor/govt entity, and in exchange they can ask questions of the bots on demand with up-to-date health info.
3
15
Show this thread
And yes, the robot (or just software) can send a fax or make a phone call if necessary.
4
17
Show this thread
At the extreme, you can imagine a robot sitting at your desk, filling out a form with a pen, or with a keyboard. It will *not* literally look like that, but the image is a demonstration of why it doesn’t particularly matter what the govt thinks about the method.
3
15
Show this thread
Most common objection in replies is govt won’t allow it or won’t integrate.
This misses the key point: bc the AI is doing the part you otherwise do, the govt side doesn’t need to change.
The govt can no more stop you from using a personal AI agent than from using a calculator.
Quote Tweet
By 2030, I expect ~everyone will have a personal AI agent that, among other things, can function as a kind of data broker to help you with:
* health providers
* service providers generally
* filing taxes
* government services generally
Show this thread
11
5
98
Show this thread
Its amazing that 10% of people still dont wear seatbelts and they account for almost half of fatalities
39
289
2,472
Show this thread
The SaaS equivalent of this, which will come much sooner, is bot <-> bot platform integrations. (As opposed to specified APIs)
Quote Tweet
By 2030, I expect ~everyone will have a personal AI agent that, among other things, can function as a kind of data broker to help you with:
* health providers
* service providers generally
* filing taxes
* government services generally
Show this thread
6
9
47
And not just personal agents. Agents for every entity:
Quote Tweet
Replying to @moskov
All government data should be within an LLM.
Immediate, accurate responses to FOIA requests.
Immediate results for any number of government services.
4
3
31
Show this thread
Filling out a tourist visa? Applying for a building permit? Complaining about a pothole?
🤖🤖🤖
2
3
36
Show this thread
Similarly, you can give your own agent, running locally, carte blanche on your financial data and have it fill out all tax forms for you. Even audits could be done that way.
W/ neat tricks like "here is a photo of a box I donated to Goodwill - what's the estimated deduction?"
2
2
32
Show this thread
Heck, maybe it can even know when fire alarm batteries are running out and send you a push notification with time to spare instead of a 100 decibel intermittent beep.
We can do it! Anything is possible in this version of the future!
3
3
44
Show this thread
The agent can monitor your home and reach out for maintenance as needed, scheduling the washing machine repair person as soon as your mutual schedules allow it.
(Don't worry fellas, you can ask it not to call the plumber w/o letting you try first.)
1
2
42
Show this thread
(Generally I think the ability for AIs to subjectively categorize actions between "I can do this", "I can't do this", and "I'm not sure/need to ask first" is an enormous superpower in itself)
2
10
86
Show this thread
Your personal AI agent doesn't need a standard though - it can just fill out the forms as they are now (maybe even on physical paper by then) and translate formats as needed.
If it's unsure or needs extra permissions from you, it will be able to ask, in your preferred way.
1
6
53
Show this thread
Filling out the same information on forms countless times every year is maddening, but slow progress on standardized electronic medical records and access protocols suggests this will still happen for the foreseeable future.
2
4
80
Show this thread
By 2030, I expect ~everyone will have a personal AI agent that, among other things, can function as a kind of data broker to help you with:
* health providers
* service providers generally
* filing taxes
* government services generally
91
146
900
Show this thread
If China deploys a dangerously advanced and strategic AI model first (at least within the next 15 years) it is very likely to be because it was trained in the US and then the model weights were hacked.
If you wouldn't trust the CCP with model X, don't train model X.
8
27
159
Show this thread
This document leaked from Google has been gaining attention. Unfortunately it's wrong and right in major ways that should make us seriously reflect on what we're creating
Yes open source models are a huge deal but more open sourcing is NOT the solution
17
66
253
Show this thread
I applaud Palmer raising awareness around this risk, which I agree is really neglected.
Unfortunately, two things can be important and neglected at the same time.
Quote Tweet
The AI Doomer argument that AI development should be paused, prevented, or airstrike'd to save humanity misses the much more likely doom vector: Bioweapons.
Perhaps superintelligent AI could do harm, but a couple bad dudes with commoditized medtech tools are so much scarier.
Show this thread
5
1
95
What the f**k, indeed, Mr. Dogg.
Quote Tweet
And now, the collected wisdom of Snoop Dogg on AI and existential risk: "Like, what the f**k?"
Show this thread
0:40
416K views
6
17
157
YIMBY policies are broadly popular and I’m tired of people pretending they’re not
35
176
1,068
90 days later & the bet is concluded. Confirming donated $500k to people in extreme poverty via GiveDirectly.org/crypto.
You can see it on chain at GiveDirect.ly/BalajiMedlock 🧵
Quote Tweet
I will take that bet.
You buy 1 BTC.
I will send $1M USD.
This is ~40:1 odds as 1 BTC is worth ~$26k.
The term is 90 days.
All we need is a mutually agreed custodian who will still be there to settle this in the event of digital dollar devaluation.
If someone knows how to do this… twitter.com/jdcmedlock/sta… Show more
Show this thread
7
138
969
Show this thread
I don't get this at all. It's not a matter of extrapolation. LLMs have *already* been way more disruptive to the technologies they compete with than self-driving vehicles ever were to traditional autos, e.g. ChatGPT/LLM search is *much* better than traditional search *now*.
137
76
846
Show this thread
And they pay for it with an investment fund dividend which is also fine and something we could do everywhere. This could be us. twitter.com/MattBruenig/st
This Tweet is unavailable.
9
9
136
I personally take Belsomra on occasion as described here
Quote Tweet
3b/ For me, the sleeping pill was a magic solution. I say "if I'm still awake in 30 mins, I'll go take the pill." 9/10 times I say that, I don't need to take it.
OTC options include Diphenhydramine (Benadryl).
I also swear by prescription Belsomra.
I don't recommend Ambien.
Show this thread
1
1
17
Show this thread
This is a great sleeping pill and I’m cautiously optimistic it will help mitigate Alzheimer’s for
millions if this research proves out in the larger study.
13
4
136
Show this thread