Navigate back to the homepage

AI for drug discovery as a Venture Capital meme - the case of Insitro

Mostapha Benhenda
March 17th, 2021 · 3 min read

In 2021, we are watching the emergence of memes as a major asset class.

  • In public markets, meme stocks like GameStop, and the SPAC mania led by Chamath Palihapitiya and Jay-Z, are running the show.
  • In crypto markets, it’s meme tokens like Dogecoin, NFT, and of course king Bitcoin.
  • In Venture Capital markets, I am seeing Insitro, an AI for drug discovery startup in the Silicon Valley, as well-positioned to acquire meme status, with a record-breaking $400 Million fundraise.

Financial memes are empty shells with little economic and technologic fundamentals. Their value is based on artistic, cultural, and sometimes even cultish reasons. They are iconic. Memes take the hype to the next level.

In 2017, enthusiasm about Bitcoin and crypto-currencies was nothing more than hype. Cryptos promised radical changes to the way we spend money, with peer-to-peer decentralized payments.

In 2021, cryptos matured into memes. They no longer promise any practical use-case. They now belong to the art, luxury and entertainment industries.

DeFi Yield Farming and Non-Fungible Tokens are the next logical stage in the history of the crypto movement. We wanted decentralized Benjamins, but all we got was decentralized Mona Lisas.

AI for drug discovery is following a similar trajectory:

In 2018, a startup called Insilico Medicine was building hype with papers promising to revolutionize drug discovery, with unlimited streams of GAN-generated molecules. At that time, I wrote on my blog how their claims were inflated, but at least, they published something to feed the conversation.

Now in 2021, Insitro (Insilico meets Invitro) keeps raising funds out of thin air, barely bothering to publish any result. With this bluff, Insitro outfunded Insilico x10.

Let’s take a look at the 3-years publication track of Insitro CEO Daphne Koller on Google Scholar:

Hardly worth $400M. Any PhD student can deliver this level of research for less than $40k.

Of course, Insitro is not in the business of publishing papers. Their job is to find cures.

However, world-class scientific research requires external feedback, from independent people. That’s what academic freedom and open science are all about.

This deep truth gets confirmed again and again. For example, last year, during the Covid-19 pandemic, the Lancet medical journal controversially published a study about Hydroxy-chloroquine, where data was fabricated by authors, and notably by Harvard professor Mandeep Mehra. This academic fraud flew under the radar of the Lancet editors, referees, World Health Organization, and several national health authorities around the world.

On the other hand, the Lancet hoax was quickly denounced on Twitter and blogs (I took a modest part in this movement, you can learn more about the full chronology of events in a previous post).

That’s why whistleblowers are protected in well-functioning democracies.

By adopting a secrecy policy, Insitro avoids the Twitter grill altogether. People at Insitro are certainly smart, with their leader Daphne Koller accumulating x3 more citations than Lancet fraudster Mandeep Mehra himself, over her career. So they know they can’t outsmart Twitter.

However, attracting top early-career talent will be challenging with so little publications. There’s little benefit for top postdocs to choose Daphne Koller’s lab, if they can’t co-author publications with her.

Apple inc. initially adopted the same strategy in machine learning, but their secrecy backfired in the research community.

Even Google, a somehow proud open research powerhouse, can’t match academic research standards, when some corporate interests are at stake. That’s one lesson from their firing of Black ethicist Timnit Gebru (which I covered in a previous post). Google couldn’t build an internal democracy with $1.4 Trillion market cap, so I doubt Insitro will be able to build an internally open corporate culture with their tighter constraints, and x1000 smaller valuation.

At the end, for Insitro, their best path is probably to become an investment fund itself, active by acquiring smaller startups, like Big Pharmas do, instead of directly getting involved into the trenches of scientific research.

Insitro might succeed as an hybrid type of investment vehicle, a kind of private SPAC for the AI era, with a clever interplay between data, AI, biotech and fintech. I actually started to explore similar ideas with ChingHub, although in a different way. If this direction is interesting for you, let’s start a conversation.

More articles from Melwy

The US-China race for supremacy in AI and biotech - Which role for the Belt and Road Initiative ?

The United States and China are fiercely competing in AI and biotech.

January 26th, 2021 · 3 min read

Black female scientist Timnit Gebru fired - the end of Google as a top AI research institution?

Ethiopian-American Timnit Gebru, a researcher in AI and ethics, got fired from Google, sending shockwaves on Twitter and beyond.

December 7th, 2020 · 3 min read
© 2019–2021 Melwy
Link to $ to $ to $ to $ to $