While I was busy explaining why the industrial research labs won’t return, an exciting thread unraveled on Twitter arguing about how to revive them anyway.

To kick things off, Activate CEO Ilan Gur argues that tech startups are the new Bell Labs, citing Moderna as proof. Then Sarah Constantin jumps in to explain that Moderna was just commercializing research from a university lab, and startups can’t do fundamental research because VCs won’t fund it.

Finally, Adam Marblestone chimes in to plug Focused Research Organizations as the missing gap between fundamental research in universities and applied science in startups.

This is all great, and highlights what I love about Twitter: smart people with diverse perspectives coming together in an ad-hoc way to discuss an important question, while making it accessible to lay audiences. [1]

I’m just not convinced it’s actually right.

Marblestone writes:

Academia’s big, so existence of fundamental research isn’t what we have to solve for (though can push more). Startups deploy well, so also not that, modulo VC risk/return params (1 of 3 proposals there). A gap is between those (another of 3 proposals highlighted).

This is a compelling narrative. Academic handles fundamental research, startups work on commercialization and deployment. What we’re missing is translational research to nurture pre-commercial research and bring it to fruition.

In fact, it’s so compelling that everyone else is already working on it. Off the top of my head, here are some mission statements to consider [4]:

• Cyclotron Road: “Our mission is to empower science innovators to advance their ideas from concept to viable first product, positioning them for broad societal impact in the long term.”
• ARPA-E: “advances high-potential, high-impact energy technologies that are too early for private-sector investment.”
• Sci-Founder: “help early career scientists start companies of their own.”

Not enough? Here’s the diagram from Actuate’s home page:

Okay, so maybe you’re convinced that the gap has been bridged, but it’s only a recent phenomenon, and only as a result of these brave pioneers.

Except here’s Breakout Labs in 2011: “Venture capital firms want research that can be quickly brought to market, and federal funding offers little room for risky, unproven ideas. We are jumping into this funding gap to energize innovative research.”

What’s going on here? Is translational research a great idea who’s time has come? An overcrowded space? A buzzword? A buzz-narrative?

Commenting on a draft of this post, Sarah writes: “I think a handful of funds existing is good validation of the idea rather than evidence it’s already overdone. The gap is probably big even if these guys are doing everything right.”

Similarly, Adam points out that the organizations listed represent a tiny share of all research funding. Gur’s thread mentions that translational research was 10% of Bell Labs’ budget. In contrast, these organizations sum up to around 1% of all federally funded research.

Maybe all of these organizations are good, and we just need to do even more. If that were true, we might expect Breakout Labs to have been an enormous success and raise/deploy more capital. But it’s always possible that while the idea in general is good, the specific implementation still matters, and Breakout Labs just happened to fail. It could also be that while 11 years sounds like a long time, it’s not quite enough to see the effects of what is explicitly pre-commercial.

Or maybe Sarah is right and venture capital, even with an explicit charter, just isn’t well suited to this type of investment.

---

Thanks to Sarah Constantin and Adam Marblestone for reviewing drafts of this post. Thanks also to Stephen Malina and rkris for their comments.

---

Appendix A: On the Basic/Translational/Applied Trichotomy

This taxonomy is worth keeping in mind, but it’s ultimately just one axis.

As Arora, Belenzon, et al. detail in section 5.1 of The Changing Structure of American Innovation, and as I summarized earlier, “inventions originating from large corporate labs are different”. Specifically, they are general purpose, practical, multi-disciplinary, and often resource intensive.

Along those lines, Google’s AI research has transformed the entire ecosystem. They:

• Pay extravagant salaries, driving the world’s best talent into AI
• Make high profile acquisitions, driving VC funding into AI
• Create Tensorflow, facilitating future AI research
• Develop TPUs, reducing the cost of compute, but only for particular applications
• Own one of the world’s largest datasets

None of this would be possible in startups or universities.

In this view, the gap is not just in “fundamental”, “applied”, or “translational” research, it’s in all the specific niches that industrial research labs could have funded, but can’t anymore. Different work is possible, and without those funding mechanisms, this work won’t happen.

That doesn’t mean the solution is a new government entity aimed at fostering “multi-disciplinary” research or whatever. We should just think about what particular aspects of Bell Labs were praiseworthy, and attempt to create the funding mechanism that will allow them to exist in some new form today.

---

After writing this, rkris and Stephen Malina told me about Spark, created at Berkeley, and pytorch, created by Soumith Chintala.

Soumith’s contributions to Pytorch seem to have occured mostly in 2016, at which point he already worked at Facebook. Pytorch has 46k stars on Github, compared to 153k on Tensorflow. Spark is at 29k.

Eyeballing the charts, the other early torch contributor as Yangqing Jia who was also at Facebook.

So I may have overstated the impossibility of creating a major ML framework in a university setting, but it does seem like Tensorflow is dominant, and Pytorch was heavily nurtured by Facebook.

Spark really was created at Berkeley. Aside from looking at the Github stars, I can’t really judge how influential it is compared to Tensorflow.

Stephen also notes that there are startups working on ASICs for AI.

Appendix B: Were mRNA vaccines invented by startups or universities?

Gur writes: “Startups are the most vibrant environment today for Bell Labs style intermixing of fundamental research and applied systems. And they are changing the world in Bell-Labs magnitude ways. Just look at @BioNTech_Group and @moderna_tx”

Sarah Constantin rebuts: “This is false. BioNTech and Moderna both commercialized/developed platforms based on innovations from academic labs.”

Who’s right?

In academia, Karikó was “issued an ultimatum, if she wanted to continue working with mRNA she would lose her prestigious faculty position, and face a substantial pay cut.” BioNTech licensed the work she had done with Drew Weissman, then hired her as Senior Vice President. She would go on to lead their work with Pfizer on the mRNA Covid vaccine. (Wired) The article adds that Karikó was considered “not of faculty quality” and that the Upenn admins laughed at her when she announced she was leaving. Another source adds that Karikó “spent the 1990s collecting rejections”.

For his part, Weisman still works in academia, though his lab is funded by BioNTech (MIT Technology Review)

And just to complete the holy academia/industry/government trinity, a third party organization claims Moderna "received approximately $20 million from the federal government in [DARPA] grants several years ago and the funds “likely” led to the creation of its vaccine technology. "

So who deserves the credit? Constantin is right that the innovation originally came from an academic lab, but this should come with the caveat that the university actively attempted to push out the lead researcher. It also seems likely that BioNTech would not have funded the research on their own, had they not seen Karikó publications.

All considered, I don’t walk away feeling like startups or universities are particularly good at fundamental research. Instead, it feels like a miracle that we ended up with commercialized mRNA vaccines at all.

It’s hard to know how much counterfactual impact government funding had, but if the DARPA grants were responsible for Moderna’s success, it’s a promising model for startups working on high-impact long-term technology, funded by someone other than VCs.

Footnotes

[1] It also highlights what I hate about Twitter. Gur provides no evidence for his claims, and is totally wrong on several of them. Constantin makes a vague complaint about VCs based on a single anecdote. [2] Adam’s tweets are confusing and difficult to parse. Everyone is talking past each other to further their own narrative. [3]

[2] Constantin writes:

A startup has to prove it can make money for investors within 5-10 years. The funding model’s not designed for long research projects… At a core level I think this is why my startup failed. I wanted to run a $1-5M experiment that nobody else was doing and seemed like the obvious first step in searching for anti-aging drugs. VC’s weren’t interested. You’re supposed to do that part in academia.

This might be true, but we should be very cautious of the argument “VCs wouldn’t fund my company working on X, so they must not be interested in the category X belongs to.” Per Thiel and Collison, startup failure is overdetermined. Maybe your startup couldn’t raise money because VCs are myopic, but there could be 100 other equally valid reasons.

[3] I’m being polite here, I really hate Twitter. Gur’s thread is filled with distracting gifs to the point of being nearly unreadable. The Constantin thread is over a dozen tweets long, with replies to individual subtweets that branch off, making the conversation impossible to follow.

I don’t know why anyone even tries to use this platform for substantial discourse. My best guess? They aren’t. Discourse was never the goal, just hot takes, dunks, and self-promotion.

I don’t say this to criticize the specific people involved, it’s all endemic to the platform, and not really anyone’s fault. Except for mine for reading the thread, and yours if you participate in this hellscape.

[4] I didn’t have to dig around to cherry-pick those quotes. It’s just the first line of each organization’s about page.

Tom Brady just won the superbowl again.

Previously, you could have attributed his streak to the Patriots as a team. But then he left for the Buccaneers, which had not won a superbowl since 2002, and immediately won again. He pulled Rob Gronkowski out of retirement who had previously been with him on the Patriots, so maybe it’s the two of them together, but still.

It’s even more insane when you remember that Tom Brady’s first Super Bowl win was in 2002. Was he some rookie back then, nominally on the team but not really contributing to the victory? Not at all, as in 2021, he was named MVP.

That makes for a 19 year rein, spanning 10 Super Bowl appearances, 7 wins, and 5 MVP awards.

It’s not just football. In other sports, we have:

• Roger Federer: First won Wimbledon in 2003, last won in 2017, and still ranked #5 in the world in 2021.
• Serena Williams: First won Wimbledon in 2002, last won in 2016, currently #11.
• Magnus Carlsen: First ranked #1 in 2010, still ranked #1 in 2021.
• Tiger Woods: First won Masters in 1997, last won in 2019. Ranked #1 for 683 weeks.

Is this weird? It should be. It’s not just that a 43 year old man is the best athlete in a full-contact sport, it’s that athletes are getting much better across the board, but still can’t supplant last decade’s champions.

Consider the progression in men’s marathon times:

Even more dramatic, here’s the progression in men’s 100 meter times:

It makes sense that humanity is getting better at sports. Presumably, we’re improving at sports nutrition, medicine, and coaching, but also just have more humans and better talent selection. Over time, each generation should be better than the last.

So why can’t anyone beat Tom Brady?

Maybe he really is just uniquely good, but you couldn’t say the same of Federer in tennis. Rafael Nadal has also been among the world’s best from 2005 (first French Open win) until today (ranked #2, won the French Open again in 2020). Djokovic, currently ranked #1, first achieved the ranking in 2011, and first won a major Open (the Australian) in 2008.

First, it’s worth asking if this is actually a recent phenomenon. Taking a look at the list of Wimbledon champions, Federer is #1 of all time with victories spanning 14 years. William Renshaw is in second with victories spanning just 8 years, followed by Pete Sampras with victories spanning 7 years. After that, it’s Djokovic again across 8 years.

The crazy thing is, it seems entirely possible that Federer could keep winning. He made it to the Wimbledon finals in 2019 where he had a close game (the longest final singles match ever) against Djokovic, and was playing well in 2020 until stepping back to recover from an injury.

What about women’s tennis? Here’s the list of Wimbledon champions. Serena Williams does not have the most wins ever, but her victories span 14 years. In contrast, the top players are Martina Navratilova (12 years), Helen Wills Moody (11 years), Dorothea Lambert Chambers (11 years), Steffi Graf (8 years). There’s one exception, Blanche Bingley won from 1886 to 1900. Not to take away from her accomplishments, but since the championship had just begun in 1884, it’s fair to assume there just wasn’t as much competition. At the very least, Williams is unmatched in modern history.

Is the continued dominance of Brady, Williams and Federer proof that we’re living in the greatest era of history, or a sign that we’re no longer improving?

In football, other players don’t even come close. Brady has MVPs spanning 19 years, compared to 8 for Joe Montana, and 4 for Eli Manning. But the award is a bit arbitrary, and not an objective measure of individual achievement.

In golf, Tiger Woods’ Masters Tournament victories span 22 years, compared to 23 for Jack Nicklaus. No other player comes close.

Finally, here’s the timeline of #1 ranked players in chess:

Carlsen’s 10 year streak is impressive, but doesn’t quite match Kasparov’s 21 years. Still, Carlsen has time. Kasparov was 43 when his reign ended and Carlsen is just 30 today

But still, the fact that any of this is happening at all is strange to me, given how much better humans have gotten at running. Maybe the advances in sports medicine lengthen athlete’s careers more than they help new players excel.

Or maybe it’s all just selection bias since I only looked at the athletes that came to mind.

In conclusion: In men tennis and football, current players have unmatched longevity, winning the top award over longer time spans than their predecessors.  In women’s tennis, it’s a slim margin. In Chess, Carlsen may continue to dominate and beat Kasparov’s record. In golf, Tiger Woods is one year behind the historical longest streak, but still stands a chance of winning again. More work is needed to see if this phenomenon holds true in other athletic competitions.

---

Simon M says:

This is one of my favourite topics, and I enjoyed your post, there’s some more ideas which I’ve pondered for a while which you didn’t touch on:

1/ Increasing economic rewards. (You earn far more as a top professional now, so staying in an extra year, or ten is much more worthwhile)

2/ Catching the wave at the “right time”. If the tools for longevity (training methods, drugs, whatever) are a recent phenomenon and there is some advantage to incumbency, then you would expect the current crop of top players to last longer but also hold off competitors for longer so even talented ones will have a shorter window where they are top of the sport.

3/ In soccer, Messi and Ronaldo have been dominating the game for a v. long time.

These are all good points. The second presents a particularly interesting dynamic. Using the toy model:

1. Athletes get better over time due to experience, but more and more slowly
2. Athletes get worse over time due to again, and this accelerates

A really interesting implication of improved “tools for longevity” is that there will never be another Tom Brady. The next greatest QB might be 30 by the time he outpaces Brady. Since 43 year old Brady is so good, the next 21 year old Brady won’t have a chance to win Super Bowl MPV.

---

Daniel Filan mentions over email that Go players are holding the #1 spot for less and less time. There’s a pretty wild chart:

Shin Jinseo comes out of nowhere and blows everyone else away, attaining the highest rating of all time by a healthy margin. XKCD provids a similar chart for chess:

Notably, the go players are also super young.

I write full-time. That’s not to say I’m at my desk 40 hours a week, just that I don’t have anything else going on.

In January, I published a post on average every other day, then didn’t publish for a week until today.

The latest post is good, but does it represent a week’s output? Why does everything take so long? Am I even trying?

There are a couple answers:

• I only publish ~50% of the posts I write
• Every post I publish has to be rewritten at least once

It doesn’t take too long to hash out a quick draft, but then I try to get feedback, make the post more compelling and easier to read, and fact check my claims. That all takes longer than the original writing process.
Not everyone is this way. Scott Alexander and Byrne Hobart (#1 and #2 in Substack’s Technology section) have claimed they write more or less stream-of-thought. Maybe it’s a magical gift, but maybe it’s because they’ve both been writing for 10 years.

Feedback is tough as well. I try to send every post to at least one person before publication to have some level of editorial accountability. Often, it’s Alexey Guzey, known for his contrarian takes and brutal criticism. This is great for readers since it improves my writing, but also means that rather than nitpicking grammar, feedback is likely to expose a foundational flaw in my reasoning, resulting in lengthy rewrites.

In other cases, feedback takes time because I want to consult the people in question. Before publishing the case against StatNews, I tried to get in contact twice. Before publishing against Lambda School, I tried to get in touch with their Chief of Staff. I also sought out feedback from Adam Marblestone and Ashish Arora before commenting on their respective publications.

Getting feedback from an actual expert is really tough. They have spent, in some cases, entire careers thinking about the topic I tried to understand in a couple days. I am often embarrassed, try really hard to not say anything wrong, and take their feedback very seriously. This all takes time.

Finally, I am not an expert in anything I write about, so there is a learning curve spanning days or weeks. Especially for posts that criticize the original source, I am very scared of making a bold condemnation of someone else’s work, and then realizing that I’m totally wrong.

On that note, I also write about a broad variety of topics, which means I’m more or less starting from scratch each time. It wouldn’t take me very long to write another post about Lambda School since I already have the required context, but I’m not that interested in writing it, and I don’t think you would want to read it. I read their recent report and have attempted to stay updated in case it turns out that I was wrong. My impression is that they are still misleading students, and I don’t have much more to say.

This blog is an attempt to learn in public in real time, which means writing about things I don’t already understand. Additionally, the deeper I go into any one vein, the more context each reader needs to understand what I’m talking about, and the less likely it is they’ll have read all relevant previous posts. Because blog readership is growing quickly, I don’t assume the median reader is familiar with my previous work.

You may worry that this means never going deep enough to get anywhere interesting. So why is this blog even worth reading?

In short, I have different incentives and opportunities. Ashish Arora is an academic, and his only way to express a professional opinion is to publish a paper. He’s not going to make the point I did about Bell Labs, even though he’s more than intellectually capable of doing so. For most of the world, blogging is still a very weird niche.

And on the other end of the spectrum, you have popular writers trying to build their following. They grow their audience by having an opinion and making bold, easy to follow proclamations of belief. Of course, there are exceptions, but you might be surprised how few. Taking blogging seriously correlates with taking Twitter seriously, which means I’m part of a vanishingly small population that does the former without the latter.