I'm a bit confused, there are already useful results from multimodal fusions? E.g., relevant to this article, there are papers demonstrating that genomic + H&E information leads to better predictions on cancer survival outcome tasks than H&E or genomic inputs alone. 

The paper you've attached implies that (some) experts already agree that multimodality is already pretty present in existing models: 

Others highlighted that AI models trained on both sequence and structure have shown promising results in bioengineering applications, suggesting that multimodal integration is not an insurmountable barrier.

Also, this is a pedantic point, but I think there is a mistake in the PDF: 

Several participants noted that significant advances in dynamic modeling are already underway. New AI-driven tools (such as BioEmu, Adaptyv Bio, and improved binding prediction models) have pushed beyond the limitations of earlier models (such as AlphaFold), suggesting that dynamic modeling is progressing rapidly (Lewis et al., 2024; Cotet et al., 2025).

As least as far as I can tell, Adaptyv Bio is not a tool/method, rather a contract research organization (CRO) that can do expression/binding affinity/thermostability assays :) it's a very good CRO and one that I have used before, but they don't seem related to dynamic modeling

Sure! Agree with that all that, including in the 250x value is very much a Youtube-optimized headline than what the episode is actually about. They also obviously study clearance antibodies as well, alongside other measures of efficacy. 

Past that, many people in the vaccine world are quite optimistic on Soham’s approach. There is indeed a trust problem in India, but smart people there are deeply aware of it and are trying to combat it. 

That's fair! I agree that a paper would be better. The counterpoint to that point is that plenty of bio startups don't prioritize peer-reviewed papers given the time investment, and that the NIH clearly finds their data trustworthy enough to fund and conduct a phase 1 trial using their vaccine. 

Yeah I can see that; I guess what I was trying to get across was that Plasimidsaurus did do a lot of cold reachout at the start (and, when they did do it, it was high-effort, thoughtful reachouts that took into account the labs needs), but largely stopped afterwords. 

Hoping to learn more about the limiting factors via people who reach out after reading this piece :)

I do somewhat buy the talent shortage + funding argument, but I also agree that there is more to the story. It may very well be the case that hundreds of these boring companies exist; I just don’t know about them. Will update this thread in a few weeks with whatever more information I learn!

That’s very true, but I do think the translation to privatization can be useful! Helps push for better UI/UX, better support, and even better infra work. This isn’t true across the board, hard to imagine a company creating something like MMSeq, but I have to imagine its true in other areas

Good CRO’s are just an example, other examples are better physical/software tooling for lab ops and better lab [object] manufacturing

Docusign does also have several other products, but that’s fair, they may very well be bloated. I do imagine that the bulk of the company is not engineering, but sales. Looking at LinkedIn confirms this: 19% engineering, 28% sales.

Yeah Arcadia is definitely less on the blue-sky/high-variance realm of the spectrum, and closer to 'better research in underserved areas of biology'. 

I was pondering adding Altos Labs here, alongside Retro Bio and Newlimit, but they do feel a bit different from others here given the strong focus on a for-profit system (Arcadia's for-profit focus is a bit more opportunistic). 

Arc probably has discovered the most influential thing: https://arcinstitute.org/news/news/bridge

Arcadia Science has a grab bag of a bunch of interesting stuff: https://research.arcadiascience.com/ 

Everyone else is a bit too new to have released many interesting things