Elizabeth Hudson retweetledi
There could be no Independence Day without Memorial Day. 🇺🇸
English
Elizabeth Hudson
5.7K posts
Elizabeth Hudson
@ClarkPolner
Founder @ Ambitious Bio. fmr @yale; industrial robotics @symbotic; Advisory Board @centerforastro; Intel @usnavy
Cambridge, MA Katılım Ağustos 2014
2.5K Takip Edilen1.2K Takipçiler
Elizabeth Hudson retweetledi
Today is our nations Memorial Day. Enjoy your day and take a moment to remember the true meaning of this day. A day to pay our respects to all those who have given their lives in our country's defense. God bless these brave heroes and their families.
English
Elizabeth Hudson retweetledi
Every year, I share this video of French caretakers who take sand from Omaha Beach in Normandy, and scrub them into the letters to give them the gold coloring.
They do this for all 9,386 US soldiers who died.
France also gave us this land as American soil. #MemorialDayWeekend
English
English
@_buggles ...generated on-demand, administered without additional testing.
Reorient medicine towards zones with strong safety priors for both targets & modalities, use improved models to cover residual uncertainty
English
Getting cliche to say now but the AI plans to “cure all human disease” fixate on the CS/tech legible parts of drug discovery which aren’t the overall bottleneck
Your awesome lead candidates will still take a decade to get approved
Generative personalized medicine is the way
English
@brad_woolf @OmicsOmicsBlog Public domain and private combined?
English
How many people have gotten their genomes sequenced globally?
Extremely conservative numbers I’ve gathered (upper end)
100M for any type of read- short, ancestry, 23&me
5M at most for WGS
Does anyone know?
English
@PatrickHeizer The retirement system part makes sense. But aren’t basal healthcare expenses pretty low? And everyone will die, but if that’s something is more acute and less chronic over decades before you die? Where did that plot above come from?
English
In addition to healthcare costs, there are the effects on Social Security and pensions.
An extra, say, 1-2 years of life expectancy with no increases in retirement age is a large hit to the financial stability of any retirement system.
English
In the long run, decent chance that GLP-1s are going to **increase** healthcare expenditures.
Preventing chronic disease means more years of basal healthcare expenses, while patients still ultimately have to die of something.
Plus the GLP-1s themselves.
Still worth it.
Sri Kosuri@srikosuri
I’m guessing it would be cheaper for the US to just buy Lilly and give all US citizens GLP1s for free than continue to drown in long term Medicare/Medicaid/healthcare costs.
English
Think about drugs like military campaigns: you need a warhead (the thing that has the effect; eg killing something). The missile - or some way to move the warhead into neighborhood of its target. And the target. The building you want to hit. All enemy; no civilians.
The first two problems are engineering problems. When you read about something cool in the news, it’s usually in those categories. The third problem is a brute force, elbow grease, data collection problem. It’s Lewis and Clark putting one foot in front of the other, heading west and drawing maps. (Drawing maps here = chewing through bodies to see of what they’re made.) But if you don’t have a good map, the rest of it is meaningless.
You’re dropping a weapon on a building, but your bombardier is drunk, has one eye closed, and is working off a map with giant holes in it. The chances you hit the enemy and nothing else are pretty small. (Dave Rick’s on Cheeky Pints said the map is 90% holes.)
English
@jrkelly any good data on what the average foot of bench face is being used for? Some combination of the sales that the OEM’s are reporting, details in CRO job postings, etc.? Looking less for what can be done and more for what IS being done (with any scale).
English
Elizabeth Hudson retweetledi
If LLMs can become lawyers and solve Erdos problems by reading the internet, should we expect them to solve disease too?
Only if the rules of human physiology is latent in our written corpus. Sadly, I don’t think they are. We haven’t measured, let alone documented, enough of how disease unfolds over time.
But we can act to make sure that data exists, instead of relying on indefinite techno-optimism. From @NornGroup’s latest blog post (quoted)
Martin Borch Jensen@MartinBJensen
English
“America will manufacture more cruise missiles in the next three years than the last 30 years combined.” ❤️
Palmer Luckey@PalmerLuckey
Many people don't understand the scale @SecWar is demanding. The United States has built less than 40,000 cruise missiles, ever, since the Army asked the Dayton-Wright Company to manufacture an Aerial Torpedo in 1917. The more time passes, the fewer we make and the more they cost. Leaders like DepSecWar Feinberg and @USWREMichael are actually responding to foreign threats and giving clear demand signal that is bringing billions of dollars of private capital into solving this problem. Now America will manufacture more cruise missiles in the next three years than the last 30 years combined, at a fraction of the cost. We are in the middle of hiring for Barracuda production at Arsenal-1 starting later this year, please hit us up if this mission is one you can be passionate about.
English
"The challenge [is] integrating data from imperfect animal models, an incomplete understanding of human biology, and in some cases no clinical precedent."
Does context engineering in physical sciences look different than with text? In a Reddit thread, does the omission of a specific word have the same meaning as a gene not showing up in a transcriptomic readout or a disease not showing up in a medical record? Interesting missingness + context-engineering + representation / priors challenge.
(Also operationally - this is why you need the machine / meta data paired with output, to interpret it, which you won't, if you outsourced it to a CRO and didn't ask for the docs at the time.)
Samarth@samarthathreya
English
“God made men differ in strength, but Sam Colt made them equal. Likewise, God made men differ in intelligence, but Sam Altman made them equal.” Beautiful
tomie@tomieinlove
@captgouda24 The prostitute comment is quite apt. God made men differ in strength, but Sam Colt made them equal. Likewise, God made men differ in intelligence, but Sam Altman made them equal. We’re now competing on beauty. The concern is not if scaling has hit a wall. It’s if you have.
English
Can you get the basic biology right without what exists today in the public domain + what privates are on their way to generating, or is something more or different needed?
Ran Blekhman@blekhman
I've said it before: the bottleneck is not speed. This DNA model is 275x faster, but it creates a tree of life that groups drosophila with plants. How is AI going to solve cancer if it can't get the most basic biology right?
English
Elizabeth Hudson retweetledi
Ukraine may have found a cheap way to give a drone extra range.
Let a balloon do the first 42 km.
ㅤ
Ukrainian troops reportedly tested launching the Ukrainian-American Hornet one-way attack drone from an aerostat. The balloon carried it 42 km and released it from 8 km altitude.
ㅤ
The drone used only about 5% of its battery before beginning its own flight.
ㅤ
The trick is simple: the balloon provides distance and altitude, while the drone keeps almost all of its battery for the real mission. If proven at scale, it could push Hornet’s estimated 100–150 km reach by another 1.5 to 2 times.
ㅤ
This is what Ukrainian defense innovation often looks like: clever ways to stretch what already exists.
ㅤ
English
Hope they’re not teaching math
Mayor Zohran Kwame Mamdani@NYCMayor
I know a few teachers in Queens who would beg to differ.
English
[Need industry to source the raw inputs for all of your automated experiments]- literal chunks of humans] - [then your model / automated lab loop goes here] - [need industry to test the outputs without taking a billion years and dollars for each one]. Not a naysayer - but I want to live forever TOO so please don’t forget about steps 1 and 3.
MTS@MTSlive
SITUATION DETECTED: Demis Hassabis said at Google I/O that solving all disease is close, as Google unveiled Gemini for Science.
English
Elizabeth Hudson retweetledi
@adamfeuerstein @matthewherper And he answered the tenure question how
English
Marc Tessier-Lavigne, CEO of Xaira Therapeutics, talking about using AI to make entirely novel (potentially) therapeutic molecules that do not exist in nature.
Wow type stuff
@matthewherper also asking about his tenure as Stanford president that was cut short (he resigned),after accusations of data manipulation surfaced.
#STATBreakthrough
English
and a multiple comparisons correction wipes out the interesting details mention / love it
English
"...Nowhere is [training data] more expensive to acquire than in pharma, where a single bit, the binary outcome of a trial, can cost billions."
Jonny@hsu_jonny
I'm super excited to announce our $7.2M seed round + a strategic collaboration with @PacBio! Together, we'll be generating long-read, bulk-RNA sequencing directly from patient tumours spanning multiple cancer indications. This is extremely rich data that will be used to further train and evaluate our models. @blankbio_, we are focused on training RNA foundation models that can learn and predict disease progression and treatment response in patients. These are the biological insights that will enable us to design smaller, more efficient clinical trials with higher success rates. Huge thank you to all of our investors who have been amazing supporters from the beginning. For more details, please check out the links below. Press release: blank.bio/seed-announcem… Blog from the team: blank.bio/accelerate-cli… We're hiring: ycombinator.com/companies/blan…
English