Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
DeSci Office Hours
Recorded: April 3, 2025 Duration: 1:02:51Space Recording
Short Summary
The discussion covers innovations in decentralized science, particularly the use of blockchain for data ownership and AI in drug discovery. Significant funding in AI for drug research is noted, alongside trends in interdisciplinary research and the push for open access in scientific tools and healthcare solutions.
Full Transcription
Music Thank you. Thank you. Hi everyone, I think we can sort of kick off, it's 5.03 where I am in Berlin, but yeah,
welcome to our office hours. This is actually something that we do
in our Discord server.
We started it as a learning exercise
for our Edu Cohort,
which is a group of about 40 people
that are currently doing a DSci course
with Molecule.
But after sort of doing a lesson
on the principles of Web3 and DSci
about transparency and openness,
we thought we should build out in the open a little bit more.
So we're having an office hours on our Twitter,
and it's going to be led by two of our favorite team members.
No shame, actually, I can't say that.
We have no favorites.
We love everyone.
Kiho and Colin, if you guys want to introduce yourself.
So, yeah, I can go ahead.
I'm Kiho.
I'm reporting from India right now.
Sorry, my voice is a little different and weird here because I think I caught some cold.
different and weird here because um i think i caught some cold but yeah um my journey into
desai started with molecule um exactly uh one year ago 2024 i didn't have a clue on what i'm going
to do next after my journey in academia and then tried entrepreneurship for some time. And then that's when D-Sci came in because it just addressed all the pain points
and the trauma that I went through in academia.
And then, you know, I'm just baptized.
From then on, I'm a D-Sci food soldier.
And this has been an incredible journey.
And especially kudos to Molecule team
and the team that I'm working with,
like exceptional people have pushed DSi
and also people into DSi.
Yeah.
And I would like to have my buddy,
Karlyn introduce himself.
Yo, yo.
Hey, how's it going, everybody?
Welcome to Office Hours.
I'm Carlin.
I am coming to you from Chicago today.
So it's about 10 a.m. for me.
Nice and sunny day. My journey into DSI, I'm probably like a couple months behind Kehoe's journey. I joined Molecule in May of 2024.
I am a researcher by trade. I'm a PhD student currently getting my PhD at Northwestern
in molecular biology. My research is a lot around like basic science, but I had sort of a drive to pursue,
you know, translational stuff, um, get more into like figuring out, you know, what goes
into all the different, um, you know, cures and treatments that biomedical, uh, research
is able to come up with, um, that led me to D-S to DSi. Really enjoying my time here. The DSi team,
the molecule team is really awesome. Yeah. And I just, you know, love talking anything science,
love talking anything Web3. I'm really still new to like the Web3 of it all. But now I'm currently
a drug dev for pump science. So if anybody wants to talk about pump science and what they got going on happy to talk about that as well but yeah with that we can kind of open our
office hours great to great to have everybody yeah so for everyone on the call um office hours are
totally free and open um just request to speak and i will grant you access and you can you can
contribute to the conversation um Um, but yeah,
we wanted to have it as like an open thing. Um,
but maybe to kick off, um,
I think we, so this week in our lessons, we covered sort of, uh,
the topic was called science 101, which obviously it's like impossible to cover
all of science in a one-hour lecture um but the
things that we really looked at were like sort of like the basic principles of like science and why
it's important to have it in our lives and like the process of it um logan who's our biotech
operations manager she gave the lecture she's on this call now um but you will only see her as the picture of her darling dog, Pippa.
She has Pipcoin.
Yeah, Pipcoin.
And then we kind of went through IP in the cost of drug development and the processes that sort
of need to happen to bring a drug to market which I think generally is
actually like quite an interesting thing whether or not you're a scientist or
even interested in it I think it's really important for all of us to have
an understanding of how medication makes it sway to us.
I actually remember a quote once that was like, everyone will be, will have a
disability at some point in their life, you know, if they age. And it's one of
those things I think it's like everyone will interact with medication, healthcare
and medical services at some point in their life, you know, for good and bad reasons.
I mean, you can be having a baby and that's a great reason.
But these are all things that sort of permeate our lives.
And it's important that we understand how these medications come to be.
But anyway, I'm rambling a little bit.
We've got a requested speaker.
How does the speaker? Wexer Hleer, I think. Just waiting for that to connect. Hey, what's up?
Wexer Hleer, we are giving you the floor because you requested to speak.
Do you want to join the convo?
Yeah, it's Vexhalia.
Vexhalia, Carlin.
It's a hyper.
Wow.
I'm so honored to be in this space.
I just came online and I saw this space.
I was like, okay, let me join in because I hosted my Desai 101 space.
That was yesterday night.
And I think we need to bring Desai to the awareness of people more because there are lots of things that's going on there i know people are not really attracted to it because it does not have the numbers you know
making 10x making 1000x immediately we are building um real world assets and solving
real world problems so it's actually something that is growing gradually but it's a long term.
And also I'm very interested in this side because it's very personal to me.
You know, someone with genetic disease and, you know, and also I would love to learn more so I can also teach people.
And, you know, this side besides changing the ways traditional science is being
run that's using the blockchain to make it more transparent to make it more community-based and
to bring it out from the hands of bureaucrats the bureaucrats yeah so i would love to learn more and
i want to know what molecule is building because i would love to have projects of my space here so we can talk
about it and also talk about um data how to make it more transparent even in healthcare sector you
know how to make um your data that it can be decentralized and you can have authority rights
you can own your data instead of it's being sold to companies to ai yeah and also
um there's this problem of maybe if you travel to another country you can't take your medical
records with you and you can't be properly uh attended to because of your data just like my
friend that traveled from one country to the other
and he wasn't really attended to very well
because he thought he was actually shaking it
because he wasn't with his medical record.
So I'm just happy to be here
and I'm listening and I'm learning and I'm connecting.
So thank you so much.
Thank you so much. Thank you so much, Vexa.
Thanks for the intro and your insight and interest
into DSi.
So happy to hear that you are interested in connecting
people around DSi.
Maybe Ella, behind Molecule, can take the first question.
But I want to take on the question that you asked about data being more transferable, transnational and being on chain. This is something that I've been also thinking and working on. So one of the projects that I'm working for, we are kind of looking at data as an asset for the users.
So I guess you know these soul-bound NFTs or soul-bound tokens.
So imagine something like that.
Whenever a data is generated from a user, be it any data, right?
Like your location, geolocation, your health data, what you purchased and all of that,
all these data comes out of us and we leave this digital trial as we move forward.
And what if we could kind of sandbox this entire data that is coming out of us
and then use this sort of a private asset?
And if someone wants to use this data, has to pay us to use the resource that is generated by us.
So with this vision, we came up with something
that's built on ocean protocols, data NFTs.
So the idea is once, let's say,
let's consider a wearable device,
you have a watch, a tracking device
that tracks your health metrics and then all this data
that you are kind of you know generating with this device ends up in a server but the access to the
server could be secured in this web3 way like private public key system and then the private
key tells that you are the owner of it you can verify and you are the owner of it, you can verify and you are the owner of it. And the public
key could then say for us, like, you know, for the provider to verify that you are the holder of that
data part specifically. And then the best part is you can invite researchers or scientists or even
these AI companies to use your data, but they cannot copy or download your data. They can only run computations on top of that,
but they cannot download this data.
And for each compute, you can monetize them
and kind of get it monetized and whitelist them
for a certain period so that they can run computation
on top of your data.
So I personally believe that privacy should be consumer focusing, like the user focused
user, because whenever you go download an app and try to use it, there is a privacy
terms and condition that you need to sign.
But you cannot bring up these terms. They come up with the terms and condition that you need to sign but you cannot like bring up these
terms they break they come up with the terms and you have to agree to it so what if you could come
up with the term and the and the service provider could agree to your own terms and conditions
and i think that will be the future and dsai definitely is doing that big part in bringing scientific data
to the decentralized forum, basically.
Any sort of health data, any sort of scientific data
that is valuable and part of this intellectual property,
I think it should enter this sort of structure
where the inventor and the owner of the data should have highest ownership and
sovereignty over the data. I ended there.
Wow, thank you so much for making that very clear and yeah I think this is where
the world needs to go to ownership of data. And you can also, like, imagine you giving out, like, your data for researchers to research
on instead of waiting for the government to, like, grant publishers or researchers and
funds to, like, go and research.
That's very, very nice.
And it makes everybody sing because you feel, feel yeah you're actually being cared for enough
Moleko are you here? okay yes yes did you want me to answer your question or do you
have another question for me no maybe you can just tell me what the molecule dow is actually building
then i can get more of course i actually want to say just as a side note to that conversation that
we're having about data tattoos on the call um is actually thinking about building something
very much along those lines of data ownership to kind
of deal with this lack of information and record transfer between hospitals and how it really slows
down efficiencies and people can't get the the health care services that they need or the best
best care because doctors can't find their records or don't have their papers or anything like that
so Ted if you do want to jump up and speak about it, then just request to speak.
But Molecule, so Molecule, I have to say, oh, awesome.
Actually, do you know what, Etet, you talk about the data ownership,
and then I'll jump into Molecule stuff, because let's just keep the data conversation going,
I think, if people are interested in it.
Okay, thank you very much,'m ella jmjm everyone um it's great to be here so uh my my name is ite tuk i'm more of a dev guy on the builder side so um to add to what um
the exalya said i have a project i'm working on. It's about, uh, a decentralized way to share
medical data. It spins from a personal perspective because, um, some years back I had a surgery and
then I changed cities. And then when I went to the new hospital, they had to do this redundant test that i had earlier did when
like i did the first surgery i was very i knew all of them but i did not have like sorry excuse me
so i did not have like access to the medical records so it was difficult for me but then i
had to do the test so after that i thought of from a builder's perspective
how can i remedy this situation so i started working on health chain which is uh it's just
like is a decentralized application that uses a smart contract and then using the smart contract
so uh they they use or the patient only gets to be registered by a verified hospital first
and then after that period every medical record the person or in charge all that has to do with
the patient the patient has full access to it in a sense that the patient can approve the hospital
to view the medical record I can approve a doctor that is vetted by an approved hospital
to view the medical record in terms of probably you have an appointment with your doctor, right?
So you can go there using smart contracts.
You can approve the doctor to be able to view and modify,
not just modify like update in terms of, let's say they do
some tests, some diagnosis, some prescriptions, so they can add all of that to the new data.
But after that period, the patient has the right to revoke that permission for that doctor or any
hospital to view the records. So it kind of like, wherever, which part, any part of the world you are,
you can easily like approve a verified hospital
to have access to your medical records, which is on chain.
So that's what I was working on.
I'm still working on it.
Yeah, that's that.
Thank you, Ella.
Amazing. amazing thank you ted for sharing um yeah yeah we're super excited i think um the dc edu court has brought together a really nice group of builders um from a really really varied
uh background so i'm really hoping to see some projects that sort of spring to life
over the next couple of weeks um and with people in the course.
And of course, you know, anyone who's really interested.
That's why I think we're also trying to open up, obviously, like these Twitter spaces and on the Molecule Discord and stuff so that, you know, we can have these big network effects of connecting people with the sort of same ideas um and you know the eagerness to build um and hopefully you know they can they can move
forward so yeah thank you so much um but to get back to i think the question which is what is
molecule what do we do um i think so maybe the first thing is that although our Twitter name is called Molecule DAO, we are not a DAO.
Unfortunately, it is just something that we haven't been able to change on our Twitter.
But we are not a DAO.
We are a company.
We are a tech company.
And what we do, and we have been doing for the past couple of years, is we are building the technology that allows science to get funded on-chain.
And it does that by sort of bringing the intellectual property around science,
and it puts that on-chain, and it allows people to directly interact with it
and govern it via intellectual property tokens.
So I think all of that is kind of just a fancy word
of saying that we're trying to bring intellectual property
into the future.
I think we are trying to find a way to empower patients,
to empower researchers, to empower funders,
and all of that so that they get to guide the progress of research
and, you know, funnel capital into the projects that they really care about.
So, yeah. Do you have any more questions?
I mean, I've got a million and one places that I can speak about.
I speak about some of the projects that have come through on Molecule,
some of the DAOs that we have
partnered with, all of that. But yeah, do you have any specific questions?
Okay, thank you so much. I would love to ask, how do we, how do people like the community,
how do we give value? How do we support the ecosystem, the desai space and the molecule down oh what a good
question um i think for me there's no one best way to support um genuinely i think it's about
getting into the communities that are in desai um there are a lot of really great communities
and they're always looking for extra hands and people to sort of um learn about
their mission and what they do uh i think molecule is a great place to start because we kind of are
as a service provider to the dsai ecosystem um because we provide our tech obviously um we are
friends with a whole bunch of other organizations in dsI so you know I think we're a good place to start. Follow our
socials. We've got a podcast
as well the DSI podcast. All of
those bits and bobs. I don't want to
self-promise too much. We've got a blog too.
Yeah we've got a
blog. Newsletter
all that stuff.
Check out the website and I think that you'll
be able to find a bunch of information on
you know the DSite space in general.
Yeah, yeah, yeah.
Actually, Colin and Kehoe helped us draft a really cool, and actually Jensu, who's on the call as well, a really cool learn page on our website.
So, like, that kind of, I think, explains, like, why D-Site was created and the principles and the philosophies.
But I think the best way to support is getting involved with the communities.
I think that's also the easiest way to support.
I think you can just interact with them.
And then if you find something that really resonates with you, you could maybe get in, you know, one of the DAO teams or something like that.
But really, I think visibility is the best gift that the average person can give DSI.
Speak about it at the next dinner party you're at.
Tell your friend.
Be like, hey, have you heard that the crypto people are trying to fix science?
And everyone will be like, I don't really understand what that means.
And then, you know, share that piece of knowledge with them I think I think
Disai is still really really young and pretty unknown I think once well once you're in the
ecosystem I think you kind of forget how few people actually know about it so yeah that's
always kind of my favorite thing is talk about it at the next dinner party you're at see if you can shock your friends okay thank you so much moleco i actually hosted a space
yesterday night around this side and i had over 31 persons attend and i think like 20 let me just
say 80 percent of the people that attended they've've not heard about this side. So we just turned it into an educational space.
We're just talking about this side and educating people about this side.
And my space will also be coming up again on Saturday.
Yeah.
And we'll also still be talking about this side,
but I really would like,
yeah,
thank you. That was my first first phase too because I'm really
passionate about Desai and you enjoyed it it went well yeah it did like the feedback I got
people were like I should host more about Desai because that was the first time they've ever, ever heard of Desai since they were in Web3.
They never knew that you could use blockchain in science and healthcare.
So that was the first time.
Do any of our listeners here not know what decentralized science is?
Maybe you can give me a reaction if you don't know and we can explain it.
Okay.
I think everyone does know. Okay, so we won't start from first principles then.
We've got our PhD gens in the house.
Hi, Teddy.
Sorry. Sorry. So when you
mentioned PhD gens, I just got a little curious um so i think we have
a couple of phd gems uh in this audience or at least i know carlin is a phd gen uh i want to
i want to i want to ask him a question um so given this um ai agentic roll up that's been happening and there was a
a lot of hype around it and these agents are now swarming into science so um this is just sort of a
a futuristic you know uh look forward question so do you think that uh ai should be allowed to
publish scientific papers um autonomously because that's been something
we heard first when ChatGPT came in and there was a full scientific article written by Chat.
So going down the line in future with this agentic swarms that's coming in, Google is doing this
co-pire, the researcher agent system so what do
you think what's your take on this it's a great question keeho i think um i think the best future
use case for the agents and stuff is going to be this like increased efficiency and analysis and like the carrying out of the experiments and coming up with new
novel, like candidates for different, for different like molecules and different like
combinations of stuff to test. Because I think that, you know, an AI agent could probably write
a patent for something pretty well. But I think that in describing, you know,
in describing and then in also sort of thinking about things in the way that you need to as a
like a PhD, as a scientist in order to, you know, come up with a new novel insight, I think that
we're still a little bit far away from these like AI agents being able to, you know, provide the
value that we would want to see on that front. But I think that, you know, provide the value that we would want to see on
that front. But I think that, you know, as of right now, they're really cool, like, they're
really cool tools. And they're really great. They're a really great way to, you know, speed it
up. And, you know, what we all want is faster results that are quality results that we can,
like, trust. And so I would more rather trust that these agents can do, you know,
they can do the very mechanical work of an experiment over and over again, and they can do
at the same time. And they can get rid of some of the like, I'm a scientist, I have superstitions
and stuff like that. Like I have to, you know, tilt my my pipette a certain way before I
inject something. Otherwise, I think it's not going to work.
But AI agents won't necessarily have those same sort of holdups. They'll just go do the data.
And then, you know, it'll be up to us still to sort of guide which way they go next, because I
think that they just don't have the capability just yet to sort of make the most important
decision on the like, you know, the criteria that we ourselves would be basing
a lot of our decisions off of. And maybe that's what's wrong with us as people, but I kind of
like how we are imperfect in that way. So I think using the tool, the utility of the agents
as something instead of letting them sort of like run rampant
will be better in the long run hope that answers your question yeah definitely uh i uh interestingly
i read a piece of article uh day before yesterday uh bill gates has written something about this
um ai revolution so he just listed three jobs that will be still surviving
after this AI revolution.
And out of that, biologists is one of the jobs
that will never be affected because we have something
that these AI agents lack, that is intuition.
So you rightly pointed out that thing.
So guys, Etet is rising his hand and Kelly Kells,
maybe Etet could go first and then Kelly.
Kelly Kells has been up for a second.
So I think Kelly, you can go ahead and ask your question
and then we'll let Etet ask his.
Oh, I think it was the other way around.
Oh yeah, sorry.
Yeah, no worries.
It looked a little different on mine.
Well, I'll just do real quick then.
Hey, guys, thanks for having me up, especially on a new account.
I do appreciate that.
I lost access to my other one.
I wanted to throw out, and here's the cannabis.
I think that Desai is what cannabis has been waiting for.
And I'm in the States and, you know, I don't know.
I am in the cannabis industry as well.
I wanted just to see, especially for cannabis genetics, do people, I've talked with some people, I know Desai is super early, all that jazz.
I know Desai is super early, all that jazz, but do you know of any specific project that
is really going into cannabis genetics at the moment?
I don't believe so.
I don't think so.
I came across a cannabis DAO long ago when I joined Desai last year but i don't think they are they're still
in the game i think uh yeah i lost touch but there was a cannabis dow but if you have a real use case
like uh if you have a a research team or you have a research question um yeah if you have a solid
project that you want to do i think i would um then ask you to you know like try it with this
different agencies which could help you spend this new dao or get this project funded or connect you
with people who could help you realize this project so maybe reach out to molecule bio xyz
yeah and see like try it try it try it maybe maybe um sideo um is something along the same lines
i'm thinking of you know treatment treatment options that are things that might be hard to
get past as like a you know something that would go through like a regulated clinical trial in the
same way um and so maybe sideo who who does with, you know, their psychedelics might be a great, like, sort of, you know, blueprint to go along the same lines.
But, you know, I, Kelly, I, as a cannabis fan myself, I really, really hear what you're saying.
I think that it could be a really great opportunity.
So if you if you want to chat more about that,
would be happy to. Yeah, no, thank you so much. I just followed you. I don't want to shill. I
don't know. I don't know you guys. You guys are new to me and I'm new to you. But I have worked
with a nonprofit organization, a cannabis nonprofit organization that they've been in business for
over 10 years, they have all of this patient data that has not been transcribed. And there's so much
stuff that can be done. But like you said, the legislation and the regulations, even just by
state, let alone looking at it on a national scale. I'm sorry, I got to take this call.
at an international scale.
I'm sorry, I gotta take this call.
Oh, no worries.
Yeah, that sounds great.
I'm gonna see if I can find out more about that data.
Cool.
Etat, you wanna go ahead and ask your question?
Okay, it's almost like a contribution.
I mean, but I really enjoyed the last question. I'm also going to do a little research, but it was more like a contribution to, I think, the agents could actually be helpful for automating stuff, automation.
Like, for instance, I think Kelly was talking about transcribing raw data.
So that's something that an AI agent could do in so much little time or so little time compared to humans.
compared to humans.
So in terms of like maybe actually sourcing the data
in terms of experimental work,
in terms of actual thinking,
I think that should be human beings doing that.
And then the AI could actually be helpful
in every other thing.
I mean, on the blockchain,
it could actually even be more powerful
because combining like the decentralized nature of the blockchain
with the ability of the AI agent it could allow for so much innovations in terms of like
if you're working like for a DAO or yeah let me say like a DAO so let's say every member of the
DAO with their wallet addresses for whatever reason the AI could easily compute whatever
agreement they have been voting a big proposal right so an AI agent could actually compute that
way faster than doing that on the blockchain so it could automate it off chain given the data of
the DAO members right so those are some of the ways I see the AI agent being very powerful.
I mean, I've played around some AI agents myself.
I have a Twitter AI agent.
It's been live tweeting about Web3 stuff.
I am currently working on a Telegram stuff.
So, yeah, I believe the AI agents are super useful,
but I don't think like when it comes to critical things that has to do with reasoning, I don't think that should be, especially when it comes to science, research and development, right? I don't think that should be left for the machines. Thank you.
Thanks, Nathana.
Thanks, Vithya.
What's that?
Well said.
I think I definitely agree with your take there,
completely agree with you.
In my opinion, these agents are sort of a tool
to compile your work or automate certain redundant tasks
that we have.
For example, literature review is something that we spend hours
before setting up a project.
So with the A-agents, you could compile all your research material
and then ask certain questions to get an idea on what experiments to design
and all of that.
But then it should be up to us to really make the decisions, call the decisions.
And one more place that I've been thinking where integration could really play hand in hand is
where combining this with the tools that we have for biotechnology or life sciences, right? Like
for biotechnology or life sciences, right?
Like the protein design tools
or bioinformatic pipelines that we have.
So these are our heavy software tools
that we use for using biotechnology.
If those tools are like, you know,
if get combined with this AI swarm,
then I think a life of a researcher
would be so, so, so easy. so easy and then yeah a person could like
automate most of the stuff and and and I think it it will be a gear shift in entire scientific
process because then you reduce the time in doing these redundant tasks because
I remember when I was in my master's, I used to do this experiment, which takes one week altogether from prepping to executing
and to see the results.
And then you have to wait for like the next week
to try something new.
So I like this process, but still, you know,
like with the such automations and stuff,
you can really accelerate science.
And I think that is something that um we should be
focusing on um thinking of these agents as tools to you know move the needle further
yeah for sure great for uh figuring out what i'm going to make for dinner, but not necessarily for,
you know, figuring out what the next disease state we need to go after is.
Jansu, I'm actually not sure. Do you want to hop on to the call? I'll invite you up as a speaker
to discuss. Jansu has just written like a whole twitter thread about AI and drug discovery
and I think it might be interesting if you share what you
learned
requests
not Jansu but close there we are Jansu, but close. There we are. Okay, yeah, Jansu, do you want to go in?
Okay, sorry, I had trouble to unmute myself. Yes, so I wrote a Twitter thread, but it is not published yet.
But in the meanwhile, there are news about some company that is doing the AI in drug research and they got funding for $600 million.
So there is really interesting things going on in that area.
Would you want me to share more details on it, Ella?
Yeah, yeah. Talk about like, I mean, what is the general consensus of, you know,
the AI models performing in drug discovery?
Do I mean, are they?
Yeah, yeah, yeah yeah tell us more okay
yes so um now the with the new ai agents and using ai in the drug research it is the the purpose is
to make the ai and direct discovery much faster and because they can screen many molecules at the same time and also they can do the individual customized drugs it is possible to make them with ai so it is possible in the next
few years we will have so many drug research becoming much faster i mean normally the clinical
trials start and when the drug is available it takes at least 10 to 15 years
and now with the ai it can be much much less time to do this especially the screening the
molecules that can be used in the drug and also like making the drug much, much cheaper, because normally it takes at least two to three billion dollars
to produce a drug.
But now it will be much, much easier and much, much cheaper.
Because AI can create new molecules.
Oh, sorry.
No, no, yeah, go ahead.
Yeah, and AI can scan millions of compounds in just a few days.
Normally, it takes like years to do that.
So everything will be much faster and much more precise, hopefully.
Like, for instance, the newest Nobel Prize has been assigned to Professor Baker.
It was Baker, right, Ella?
Yes, yes, Dr. Baker.
Yes, so his work unlocks the protein structures
and paving the way for much precise targeted drugs.
And it can also optimize the existing treatments by analyzing the
clinical data to predict the patient responses and also can help to adjust the correct doses of the
drugs that's awesome i think that optimization is probably going to be like the the thing that
changes the game for a lot of different treatments.
Because like, you know, as it is now, there's just not enough data to know what each person's optimal, you know, treatment plan should be.
But with like, you know, the AI power and like personalized medicine that we're getting now, I just think that, you know, there's probably a drug that already exists for a certain treatment
that someone thought didn't work for them, but like they just weren't getting the right dose or
the right combination of treatments or something like that. So the opportunities are really like,
really exciting to me as well. Yeah, I came across like various companies that are analyzing the genetic data of patients and medical history and then
they can create a personalized treatment plans which would enhance of course the effectiveness
of the therapies and also the existing therapies. Yeah once we get that plus ATT's you know
transferable medical records we're going gonna be in a good spot.
Exactly, I'm looking forward to that.
Yeah, and also one thing to be cautious of, right, like an AI model is as good as its data,
so the quality of the data that we are introducing or training the EZA model is very important.
So we all know being in the scientific domain for years,
there is, you know, you cannot rely on all the scientific papers
that's coming out.
You cannot say that's the absolute truth.
So there are a lot of different stages where bias could come in.
And so there is a reproducibility crisis.
There is no negative data shared within the system.
So I think there are a lot of issues.
So once we go take those big steps forward,
we also might bring in humans into the loop
to evaluate the quality of the data and
and and that's something very important for training these AI models so I see a
lot of hands raised let's go by order right I think Renu came first and and
then Vexa and then Akvers so go Renu yeah I actually have a similar opinion with regard to the way AI can impact.
In fact, I feel the biggest potential that AI would have would be in interdisciplinary research.
Because prior to coming into this entire environment, I was working in this organ-on-chip modeling with tissues.
this organ-on-chip modeling with tissues and over there they were trying to upscale the entire
process of personalized testing because they're trying to build these specific autonomous organ-on-chip
models from individual people like who are working stem cells from an individual person and trying to grow organs from them and they wanted to test this on like
a large scale so maybe a thousand thousand five hundred chips so of different people but testing
specific drugs and vaccines so for which they have actually built an entire robot system
for which they can integrate ai as well because then they just have to mention that these
are the compounds which have to be tested. And then there's an entire upscaling process that can
be done just with the help of AI. And then the entire analysis and the statistics can be
developed. So I see that there's a huge potential, even with interdisciplinary research,
because just sitting and working in the lab is
something different but maybe coordinating from a life science to a mechanical engineer or maybe a
computer science engineer so this kind of a cross-cultural integration would have a huge
amount of impact instead of departmentalizing a lot of the way things work. And also, I think this is where even an example of
DeepMind comes in, like Google's DeepMind, where they have AlphaFold. So I was also working
quite a lot in protein modeling, and I know how difficult the whole process is. Proteins collapse
very quickly. It's very difficult to model their actual structure, but AlphaFold was able to
provide us with a platform where this can be done. So I think there's a huge impact with the way AI
can work around and it's the right way. I think we have to directionalize it, but yeah, it's wonderful.
Rina, if you had to pick, what would be the direction that you would point DSI toward
right now?
I would actually want DSI also not just to stick with life science-based projects.
We should also see where interdisciplinary kind of research would be
possible because this is something that hasn't been explored in the traditional scientific
areas. I mean, of course, we have astrobiology, where like space biology, and then we have other
fields, but there are also specific sectors where we can harness the potential. For example,
if someone's like a nuclear engineer engineer and then we want to do something
about radioactive microorganisms.
So this is like two different labs
which are completely working on two different things,
but it can be integrated in ways
that maybe a normal person cannot imagine.
But I think AI can decipher
where exactly this overlap would be possible.
So I think that's something that I would love to also explore because it's a beautiful way to think around things.
I think that's great. Yeah. I think our best ideas come from when we work with others. So
like collaboration is key. That's awesome. Great answer.
That's awesome. Great answer.
Okay. Actually, I wanted to put in something to end.
Kansu talked about personalized data to enhance therapy.
I think there are people that are having the same disease,
but no two persons react to the same drugs at the same rate.
So I think if they could integrate AI into patient records, so they could actually have personalized experience with them.
So they know this is what happens when this person takes this.
Okay, let me take myself as a
case study like being with sickle cell there are my friends that when they take morphine it doesn't
numb their pain but there are others when they take morphine it numbs their pain so
and it's as if we have like different things that trigger us but we have like a general symptom yeah general triggers but we also have
personal triggers we also have personal drugs that are good for our system so i think if the
integrates ai it also helps in personalized um data experience so it can enhance the therapy
sections and the drug sessions and also i wanted i saw
reno welcome to the space and i went through your profile and i saw stem cell researcher so is there
any stem cell research on sickle cell or anything like genetic yeah thank you so much. I would love to learn about that.
I'm sure like there is a lot which is definitely going on, but my field of research in stem cell was actually in personalized medicine. So there is an opportunity where we did work on sickle cell as well in tissues.
I don't know if you've heard of this company called tissues, KMBH.
So they work on organ-on-chip models. So where we obtain stem cells from specific autonomous
patients, like it's an autologous system. So where the stem cells are from one patient,
but we develop organs. So like we build kidney cells, liver cells and lymph nodes, and then we integrate it on one single chip.
And then the testing is done to see if there are any changes with regard to the immune cell count or if there is any specific changes.
So, of course, the end reality would be to have an entire human on chip model, which is what they're working towards, which is really interesting.
on chip model, which is what they're working towards, which is really interesting. But in my
project, we worked on the liver, lung and nymph node in order to test vaccines for tuberculosis.
So I think there are several other labs which are working on sickle cell as well, which is pretty
interesting. It depends upon the models and the kind of specific patients that we can collect the
stem cells from, but I'm sure it's
possible and it is a very nascent stage because stem cells are still like building up research
abilities and we have failed a couple of times especially while programming normal cells stem
cells into like liver cells in itself so I think there's a lot of potential in this area. And it would work a long
way, especially in personalized medicines, because it's like a world where you can think of every
person you own your own chip. And then you can just test all your vaccines, your medicines,
all with your own chip. So you don't even have to inject anything inside your body, you can even
test whether what would happen if you
have tuberculosis may maybe at like some point in your life and how your cells would react so
this is something that's wonderful that they're really working towards so yeah
wow i love you already right now yeah. Yeah, you are actually a hero.
Thank you so much.
Hey, Aquish, you could go ahead.
You have your hand raised for a long time.
Well, thank you very much.
Can you hear me okay?
Yeah.
Yeah. Yeah.
Okay, great. Yeah. So, I mean, really interesting, really interesting stuff that everybody's been speaking of.
I, you know, I'm a little bit newer to the world of AI really impactful tools, AI tools that are, you know, so helpful to, you know, not only science, but AI companies or AI tools that are being, I guess, more so like sold in the Web2 space or being sold to institutions or being sold as a tool, how do we incentivize them to either open source or enable the tool
to be utilized in a more decentralized way because of its potential impact in solving
reproducibility crisis or verifying data or other types of things that we are trying to solve in decentralized science.
I see a future where there'll be open access
to all these tools,
but open sourcing the code that someone developed,
and there is a financial,
so there is a venture interest attached to it.
I don't think that's happening in the near future anytime soon with any of these venture driven firms.
But definitely it will be open access because the larger the data sets,
the more the people use these different models, they train and they get better.
So I think access will not be an issue, but I think open source should be the next future.
So the way how you train your models in a decentralized way, I think you might have
heard of Prime Intellect, which is doing the sort of decentralized AI model training.
which is doing the sort of decentralized AI model training.
And I think in near future, all these agents that we use in a decentralized future will have a wallet.
And then, yeah, they'll be incentivized to work for you.
And there'll be a queue to, you know, use these agents.
um so yeah that's my take on this yeah and i think you know getting toward it's gonna it's
So, yeah, that's my take on this.
gonna be hard to get people who have already made a lot of money on some stuff like this to you know
move it to an open open forum so i think maybe we need to look towards things like you know deep
seek and and all that stuff where they be they are at the very least at the very moment,
like, you know, saying that they're building these things with an open source sort of vision
in mind. And so, you know, when we put our, our support and our, you know, the, the thing that
makes this community work is the community. So when we put our people behind, you know, supporting
the things that we want to see in the future, like the, you know, open resources, things that anybody can have access to, you know, the GitHubs that don't have like a paywall to get the like source code and stuff.
That's when, you know, we'llhoe was saying, it probably won't be any time in the near future.
But I think if we follow that path and, like, you know, put your time and energy toward the ones that are doing what you like to see, we'll probably get it a little bit more on the question of open access, then, do you, I mean, the
people obviously are putting a lot of time and effort into developing these useful tools.
How do you see, I mean, obviously, more people that are using it, as Kihon mentioned, you
know, the better trained AI model can be.
How do you see open access and I guess fairly compensating these people utilizing like more, you know, Web3 blockchain technology?
Um, I think, I think there are different, uh, it, it's just the, um, the ethos
of the developer at the end
and the tool that they are developing
it for the use case and stuff.
If you see Deep Seek,
as Carlin mentioned, which is an open source
AI, you don't even need Internet to host Deep Seek.
I guess you all
you can locally host all these ai if you have the enough cpu and gpu
capacity you can download these models into your pc and then host them locally so without even
having internet connected to it so there are ways that um you know it it definitely definitely comes to the ethos of the developer
and the developing team and the use case but um there are certain areas i think we need a lot
more open access than controlled access with the agent system so one thing could be healthcare
because you know healthcare is such a such a thing that everyone needs, and it should
be everyone's right to have better access to health. And so in that case, whatever solutions
that are coming, if you build incentive structures in a way that all these developers who put their
effort in building these tools could be incentivized long term for keeping the system more
you know robust and rigorous i think they will definitely build on this i think the entire web
ethos is built on this you know deac decentralized acceleration forward so i think um that's that's definitely happening and and we also have history of things
open sourcing uh software like linux when first linux came into the picture so all this really
shows us like you know how um open sourcing could really open up the field and bring in a lot of
innovation so there is a upside to it but it definitely
definitely depends on the case right like if it is for banking or any sort of
financial sectors I think all these AI tools for website development and stuff
like that definitely they'll run on run by monetizing and there will be no open
access to it so it should be use case basis.
And if you are interested, right?
Like if you want to fund a decentralized science
or an AI project that you think,
then I think you should like, you know,
like really push it forward for that.
Like bring the developers and build these pilot projects
and show the world that these things could really work.
Because with AI, now anything is possible, in my opinion.
You can accelerate all the development so fast.
There is Replit AI, Bolt AI, which could build an app in three hours.
I have tried almost all these tools.
So I think there's a lot of potential.
I think there's a lot of potential.
But unfortunately for this call, we have hit our time point.
So I'm giving the mics to Carlin and then Ella to close the session.
But we can take this conversation offline or maybe join us for the next e-site office hours next week same time
here at twitter or x we can discuss this further or maybe something different well thank you keo
i appreciate that yeah great great questions all around um you know if you have any further
questions pop them in the discord if you're in there, you know, feel free to follow
and, you know, at me if you have anything as well on here. I'm not the most active on a day to day,
but, you know, I try to at least log in every other day. So I'll get to you when I can. But
yeah, thanks everybody for coming. Hopefully we can continue to do this every week and
people will show up and keep the vibes going
like they did today it was great thanks yeah i'll close out just to say thank you so much to our two
amazing hosts keo and carlin um yeah i really enjoyed the conversations today i love i love
that like natural conversation points just kind of float up and it's really cool to hear what everyone's interested in.
But yeah, had a great time.
I definitely think we should be doing more of these
office hours on Twitter, so expect to
see more of us.
And to everyone out there, stay safe. Have a
beautiful day, morning, evening, wherever
you are. And yeah, let's
keep on decentralizing science.
Okay, cool guys. Bye.
Bye. Thank you so much. Okay, cool, guys. Bye. Bye.
Thank you so much.
Bye, Ella. Thank you.
