I almost forgot there were ML models besides LLMs! [RFdiffusion](https://www.bakerlab.org/2023/07/11/diffusion-model-for-prot...), used for the protein discovery, is a model somewhat related to diffusion-based image generation. So much of the attention right now is about what LLMs can do with their coding, agentic, and research capabilities as they get more and more powerful and self-improving. It's fun to be reminded that ground-breaking things still can come from "your dad's neural networks."
Also... just because I was checking through the results of the paper to see if it was actually an interesting result: "The SHRT binder provided complete protection (100%) to mice . . . . The LNG binder exhibited comparable efficacy, completely neutralizing α‐cobratoxin but not the non-target ScNtx"
Yeah David Baker and co. are using a lot of diffusion models for their research!
And yeah some of these mice results are very good, and hopefully they translate to a good degree to humans
The "over 100,000 deaths" seems like a slight exaggeration, this paper bounds the estimate at 40,000-80,000: https://www.nature.com/articles/s41467-022-33627-9 Wikipedia's "Epidemiology of snakebites" had similar estimates.
That's still a lot of people! But I was skeptical of 100,000, especially without a source.
Yes you're right! It's very hard to estimate how many people die from snake bite, for a thousand reasons (under reporting, incomplete snake databases, etc). This meta analysis indeed puts the numbers above 100.000 https://pmc.ncbi.nlm.nih.gov/articles/PMC11020954/ but I get where you're coming from! The uncertainty in these data is huge.
Now, how to prevent the attempts of big pharma to patent these to make tens of thousand of Indians/Africans pay big $$.
These AI modelers are reputedly not validly a human creation = patent/copyright. These protein constructs might have a long room temp shelf life = needed in most Indian/African bite victims.
These will probably be ignored rather than patented/monopolized by Big Pharma. A few million treatable cases per year globally, concentrated in low-income countries, is a small market to chase. Medical needs in low-income countries need to be addressed by governments or nonprofits rather than for-profit companies.
One hopes a storable and long lasting product will be able to save many of those 100,000 deaths(??), but you are correct - it might not offer enough $$ to USA pharma?
Yeah there isn't enough money in it for big pharma. Maybe some other form of funding, but this is the reason why all these tropical diseases are neglected.
I'm not a chemical warfare expert, but I feel like we're probably not limited by our own ability to come up with new ideas in that regard. I imagine we already have plenty of horrible options.
Spider bites are far less deadly. Worldwide, yearly deaths from spider bites number in the dozens, not tens of thousands.
Spiders target small prey and it's hard for their venom to kill something as big as a person. It's certainly unpleasant and can cause significant damage even when it doesn't kill, but snakes are responsible for far more deaths.
> So, I love snakes: I think they are the coolest animals out there
No they are not. Said almost no-one.
These creatures contribute almost nothing to society and even if they're used to get rid of pests, there are other animals that do it better without being an extreme danger to humans or they can be safely caught and released.
But instead, these creatures just want to bite, spit or strangle you if they feel like it and in some cases really just want to kill you (because they think you will kill them first).
You cannot reason with these creatures with full of venom in their fangs even if the situation was an accident.
We were talking to the reptile caretaker at the Abu Dhabi zoo and we learnt that
- snakes are actually cuties. And more like babies. They do not have strong fight or flight ability like mammals. If they get hurt, they go into a corner and just give up
- they're extremely misunderstood beings. When a snake is hissing, it is because it it actually very very frightened.
- their eye sight is very very bad. They detect mostly by that body heat. And they've a very hard time detecting friends vaa foes. A ball python actually curls into a ball when tensed up. The caretaker had to pet the snake to help it relax.
I am incredibly afraid of snakes, and living in Australia that’s not an unfounded fear. Yet even I recognise this is an abhorrent and arrogant take. Nature does not need to serve us to earn its existence.
That is interesting and I think it's a popular opinion. I think in line with the Epic of Gilgamesh or the Old Testament.
But allow me present another opinion.
Snakes don't do things they don't have to do. They don't kill for no reason (cf humans, cats). They do what they have to do, without much emotion or fanfare.
They shed their skin and don't carry extra weight.
In other cultures, a snake is a symbol of wisdom, rebirth and healing (eg the rod of Asclepius).
Outside the symbolic arguments, I think snake venom does have some research use (eg blood thinners) and some would argue snakes are mechanically cool.
Just another perspecive ofc, and I respect yours.
In the end, snakes want to survive in a competitive world.
I almost forgot there were ML models besides LLMs! [RFdiffusion](https://www.bakerlab.org/2023/07/11/diffusion-model-for-prot...), used for the protein discovery, is a model somewhat related to diffusion-based image generation. So much of the attention right now is about what LLMs can do with their coding, agentic, and research capabilities as they get more and more powerful and self-improving. It's fun to be reminded that ground-breaking things still can come from "your dad's neural networks."
Also... just because I was checking through the results of the paper to see if it was actually an interesting result: "The SHRT binder provided complete protection (100%) to mice . . . . The LNG binder exhibited comparable efficacy, completely neutralizing α‐cobratoxin but not the non-target ScNtx"
Yeah David Baker and co. are using a lot of diffusion models for their research! And yeah some of these mice results are very good, and hopefully they translate to a good degree to humans
The "over 100,000 deaths" seems like a slight exaggeration, this paper bounds the estimate at 40,000-80,000: https://www.nature.com/articles/s41467-022-33627-9 Wikipedia's "Epidemiology of snakebites" had similar estimates.
That's still a lot of people! But I was skeptical of 100,000, especially without a source.
Yes you're right! It's very hard to estimate how many people die from snake bite, for a thousand reasons (under reporting, incomplete snake databases, etc). This meta analysis indeed puts the numbers above 100.000 https://pmc.ncbi.nlm.nih.gov/articles/PMC11020954/ but I get where you're coming from! The uncertainty in these data is huge.
[dead]
Now, how to prevent the attempts of big pharma to patent these to make tens of thousand of Indians/Africans pay big $$. These AI modelers are reputedly not validly a human creation = patent/copyright. These protein constructs might have a long room temp shelf life = needed in most Indian/African bite victims.
These will probably be ignored rather than patented/monopolized by Big Pharma. A few million treatable cases per year globally, concentrated in low-income countries, is a small market to chase. Medical needs in low-income countries need to be addressed by governments or nonprofits rather than for-profit companies.
EDIT: replaced "100,000 treatable cases" with "a few million treatable cases" per https://pmc.ncbi.nlm.nih.gov/articles/PMC11020954/ referenced by Lol_arco.
One hopes a storable and long lasting product will be able to save many of those 100,000 deaths(??), but you are correct - it might not offer enough $$ to USA pharma?
Yeah there isn't enough money in it for big pharma. Maybe some other form of funding, but this is the reason why all these tropical diseases are neglected.
This means they also are amazing at designing venoms. Biological and chemical warfare in the ai age have the potential to be really fun and amusing
I'm not a chemical warfare expert, but I feel like we're probably not limited by our own ability to come up with new ideas in that regard. I imagine we already have plenty of horrible options.
As opposed to before the AI age when biological and chemical warfare were a friendly affair?
Now do spiders.
Spider bites are far less deadly. Worldwide, yearly deaths from spider bites number in the dozens, not tens of thousands.
Spiders target small prey and it's hard for their venom to kill something as big as a person. It's certainly unpleasant and can cause significant damage even when it doesn't kill, but snakes are responsible for far more deaths.
Scorpions would be a much better candidate than spiders. They kill fewer people globally than snakes, but still thousands a year.
And ticks.
That's not quite the same. Ticks can spread bacteria, viruses and pathogens. Venom consists of proteins. I assume this impacts development.
Tick vaccines for different diseases such as lyme disease are in development though. For TBE they already exist.
Anything that effectively deters or kills them without being dangerous to humans would work for me (and is convenient enough).
Gene drive perhaps?
Can you elaborate on that? Thanks!
> So, I love snakes: I think they are the coolest animals out there
No they are not. Said almost no-one.
These creatures contribute almost nothing to society and even if they're used to get rid of pests, there are other animals that do it better without being an extreme danger to humans or they can be safely caught and released.
But instead, these creatures just want to bite, spit or strangle you if they feel like it and in some cases really just want to kill you (because they think you will kill them first).
You cannot reason with these creatures with full of venom in their fangs even if the situation was an accident.
> These creatures contribute almost nothing to society
Animals have to contribute to society to be useful to the ecology of our planet?
> You cannot reason with these creatures with full of venom in their fangs even if the situation was an accident.
You can't really reason with any creature other than a human, and that's often spotty too!
Snakes are cuties and almost all of them are completely harmless. A great many of them have very sweet and gentle personalities.
We were talking to the reptile caretaker at the Abu Dhabi zoo and we learnt that
- snakes are actually cuties. And more like babies. They do not have strong fight or flight ability like mammals. If they get hurt, they go into a corner and just give up
- they're extremely misunderstood beings. When a snake is hissing, it is because it it actually very very frightened.
- their eye sight is very very bad. They detect mostly by that body heat. And they've a very hard time detecting friends vaa foes. A ball python actually curls into a ball when tensed up. The caretaker had to pet the snake to help it relax.
- most of that snakes are non venomous.
I am incredibly afraid of snakes, and living in Australia that’s not an unfounded fear. Yet even I recognise this is an abhorrent and arrogant take. Nature does not need to serve us to earn its existence.
That is interesting and I think it's a popular opinion. I think in line with the Epic of Gilgamesh or the Old Testament.
But allow me present another opinion.
Snakes don't do things they don't have to do. They don't kill for no reason (cf humans, cats). They do what they have to do, without much emotion or fanfare. They shed their skin and don't carry extra weight. In other cultures, a snake is a symbol of wisdom, rebirth and healing (eg the rod of Asclepius).
Outside the symbolic arguments, I think snake venom does have some research use (eg blood thinners) and some would argue snakes are mechanically cool.
Just another perspecive ofc, and I respect yours.
In the end, snakes want to survive in a competitive world.