Michael F

VP Product @Iterative_Hlth, clinical trials + data + ml

[Michael F @michaelfives]

totally agree with the spirit of this. but the fact is that some patients also choose to join trials because, unfortunately, they can't access (ie. afford) standard of care in the first place - often because their insurance denied it or the copay is too insane. this is especially true for studies that supply soc as the comparator, and highlights some of the subtle ways that trial design drives enrollment.

[Dave A Ricks @DaveARicks]

I’ll take that bet.

[Michael F @michaelfives]

Positive development I think from @FDA … leaning into predictability. The tradeoff is flexibility, which matters as new programs emerge, but tbh consistency probably matters more for speed today.

https://t.co/TO6TfQ0KYL

[Michael F @michaelfives]

Ultragenyx selling a priority review voucher for $210m. This is almost double PRV prices from three years ago.

Also a clear price signal about how pharma values time to market. $210m for four months, or $1.75m per DAY for accelerated review!

https://t.co/GTHGZwjLq4

[Michael F @michaelfives]

Interesting announcement from @FDA re formation of a new office for clinical trial modernization within CDER. Particularly interesting that they are putting AI alongside RWE ... signals that FDA is building the capability to scale alternative evidence generation methods.
https://t.co/Rnsk2fwVHv

[Michael F @michaelfives]

@EricTopol @ScienceMagazine Yes. This is effectively what we’re seeing from @InSilicoMeds who designed their recent IPF 2A trial to enable them to qualify the anti-aging bio marker. Expect we’ll see them pre-specify that marker next time, which is the right pathway for evidence-based approvals.

[Michael F @michaelfives]

@DrSamuelBHume Yep! This is the playbook that has worked really well for breast cancer in I-SPY 2. Exciting to see more of these platforms being put int place!

[Michael F @michaelfives]

Nice! This is the way. Looks like it is inspired by I-SPY 2 or similar, which is a really good thing.

Good stuff from @ARPA_H

[ARPA-H @ARPA_H]

Clinical trials still move like it’s 1995.

A decade of waiting, $1–2 billion spent, and a 90%+ failure rate on a clunky, fragmented system.

ARPA-H’s SURPASS program is rethinking the entire clinical trial model.

Adaptive platforms, predictive models, real-time analysis, and AI will make trials faster and cheaper so treatments reach patients years earlier.

Ready to stop waiting? Learn more about SURPASS ➡️https://t.co/yhVDUgns4c

[Michael F @michaelfives]

This is interesting bc, along with @FDA Trailblazer (which should get more positive attention) we’re seeing the start of a trend.

Governments are figuring out that clinical trial activation speed is industrial policy, and jurisdictions are going to start to compete for studies on startup (and eventually filing) time. Will impact site selection, network composition, etc…

https://t.co/iGAXQhwNAv

[Michael F @michaelfives]

Strict scoring punishes bad execution and rewards the best molecule. We should take note of who’s asking for it.

Rigor here is a public good, and there’s a lot of room to become much more sophisticated in terms of how we ensure that trials are rigorously designed and executed.

[Dave A Ricks @DaveARicks]

EASD in 2 weeks! A lot of weight-loss numbers are about to drop..including 3 major studies from @EliLillyandCo.

Before we understand a molecule’s performance we need to understand design choices that move a topline weight loss number:
- Baseline & mix. BMI, diabetes status, sex and ethnicity all shift observed response
-Stats methods. Treatment-policy and efficacy estimands answer different questions and trial drop out matters for real world outcomes. Mixing them is not appropriate. 
- Uncontrolled tails. Open-label extensions and crossovers can keep weight falling after randomization ends. That is not a new efficacy estimate.
- Execution. Lifestyle intensity, titration schedule, time at maintenance dose, and placebo-arm drop-in of approved medicines (important new trend..see ADA survo data).

Head-to-head data settle all this and should be much more widely used in the future. Where they don't exist, an indirect comparison is one data point — anchored, covariate-adjusted, prespecified, assumptions stated.
Those rules apply to our work too. Hold us to them.
Methods first. Then the number.

[Michael F @michaelfives]

@KDPomp …and guess which team chose Ben Cherington over Matt Arnold before Arnold went on to become GM of the Brewers?

[Jacob Trefethen @JacobTref]

At the OpenAI Foundation, we believe breakthroughs against disease will come from pairing intelligence with more observations of the world – in other words, more data.

Today, we're announcing over $125M in grants, launching our second science program:
https://t.co/otD2TRJCyu

[Michael F @michaelfives]

Insilico is going to be interesting to watch. Their success in the ipf 2a is great to see, but it’s more than ipf. They designed this program around anti-aging also (which is not recognized as a disease) then used the ipf trial as a chassis …banked the blood draws then got post-trial consent to run the different aging clocks on them after the fact. They’re laying the groundwork here to qualify the anti-aging bio marker next time when they’ll pre-specify it. This is an intentional strategy. We should expect to see another trial in non-IPF aimed at that biomarker in the future.

https://t.co/J7tFhILfko

[Michael F @michaelfives]

This is an important article from @RuxandraTeslo and I’m glad to see it bringing this conversation to a wider audience. Everyone should read this!

I think the point about research shifting from the US is particularly important and bears repeating. While it’s definitely true that Australia is an important part of the early research ecosystem, the fact is that pure population scale and production capabilities means that the center of gravity is shifting more and more towards China.

Thing is, this is probably a good thing … to a point. Humanity will be better off with a strong research ecosystem in China AND the US (and across the wider world). More research capacity means better treatments reach more patients faster. But capacity isn't the only parameter. The reality is that clinical research is not monolithic. Differences in standards of care, populations, clinical practice, and patient preferences are inherently baked in to the machinery of a study. This can be …confounding (especially for sponsors), but I think it’s a feature not a bug. It means that we generate more and better evidence, which in turn leads to better interpretations, more questions, and novel ideas.

But what will happen as the clinical research value proposition increasingly shifts towards China is that the infrastructure for running trials will shift along with it. Clinical trials are fantastically complicated coordination and logistical problems, and there’s a tremendous amount of insitutional knowledge that goes into making it possible to run them. Once the network of sites, investigators, coordinators, manufacturers, labs etc… evaporates, it will be hard to reconstruct. We’ve seen this movie before. It’s the reason the US struggles to manufacture batteries, drones, solar panels, etc… despite their strategic importance and the apparent domestic will. The cold (re-)start problem is a real thing. And we should expect FiH to POC to be the first step. We’re already seeing this movement in things like Lilly’s recent agreement with Innovent for development through phase 2.

History repeating in this case is not a system-optimal outcome. The overall ecosystem benefits from a plurality of participants, not one single country becoming the monopoly exporter. In this case, that net loss is to the overall system and not just the US’s strategic interests. It’s true that US patients will lose early access to groundbreaking therapies, but patients everywhere will suffer from a less representative and ultimately less capable research ecosystem.

[Ruxandra Teslo 🧬 @RuxandraTeslo]

We are on the cusp of a wave of new therapies for some of the worst diseases. But the world won’t benefit unless the US fixes its drug regulatory system. My new essay for @nytimes, on how slow clinical trials are now the biggest obstacle to curing cancer.

https://t.co/a5qf9bAKXL

- I interviewed dozens of researchers, especially oncologists at leading U.S. centers. A striking consensus emerged: science is no longer the main bottleneck to new cancer drugs. It is our ability to test discoveries in patients through clinical trials.

- The cost of starting a Phase 1 trial in America has roughly doubled over the past decade. As a result, companies increasingly take early trials abroad: Australia’s Phase 1 trial volume has nearly doubled in a decade, driven mainly by U.S. companies.

- Unfortunately, the underlying incentives are badly asymmetric: Institutions can be blamed for harms caused by moving too fast, but almost no one is blamed when patients deteriorate during avoidable delays. One doctor called the emerging system “ritualized safety over actual risk assessment.” Or as, @DavidHongMD put it: "We often forget that the biggest risk is the cancer itself."

- This problem is becoming more urgent because medicine itself is changing. Sequencing, biological engineering and A.I. make increasingly personalized therapies possible. But our regulatory system was mostly built for standardized drugs tested in large populations.

- @sytse, the co-founder of GitLab, shows what personalized medicine can achieve: after relapsed osteosarcoma and being told there were no options left, he pursued a highly individualized approach and has now been cancer-free for a year. But doing so required extraordinary resources and regulatory expertise.

- Pierce Ogden’s father was less lucky. After molecular analysis identified a drug that might target his glioblastoma, the manufacturer agreed to provide it. But administrative barriers delayed access until it was too late. “My dad was ready to try anything,” Pierce told me. “But the system is paternalistic.”

- The A.I. revolution is making this bottleneck more important, not less. A.I. relies on relevant data. Information from early-stage trials could compound with A.I. tools to achieve truly revolutionary medicines. Without the data, this is far less likely to happen.

- Another important shift is that innovation increasingly comes from academic labs and small biotech companies rather than Big Pharma. These small companies find it far harder to unable to absorb delays and regulatory barriers.

- Apart from cancer, China is the biggest winner from America's outdated medical regulations. China has a much faster trial system, with testing often starting a full year earlier. This allows Chinese pharmaceutical companies to experiment and improve medicines while American companies play with mice. Today, half of all drugs licensed by major pharmaceutical companies originate there, up from less than 5 percent only a decade ago.

- But we don't need to copy China. The best model is Australia: lots of on-site scientific and ethics reviews, and requirements that are proportionate to small, early-stage trials. Phase 1 studies there begin roughly 6–12 months faster, without any notable increases in adverse safety events.

- Operation TrialBlazer, a 2026 HHS initiative is a good start in this direction, but we need legislative action by Congress to truly make Phase I trials faster and more efficient!

I want to thank everyone who helped me with this article: everyone I interviewed and the amazing editors at the Times. This is the result of a months long journey of extensive interviews and research.

Special thanks go to those who came on the record. One of the features of the system is an atmosphere of fear, where practitioners are afraid to publicly come out and explain these issues. So anyone who does is a hero in my book!

[Michael F @michaelfives]

Platforms like X will probably start to offer Pangram-like filters for AI-generated content. Similar to what we’ve already seen from Substack and LinkedIn.

This will put evolutionary pressure on AI “slop” style. Red queen dynamics mean that automated “content creators” will navigate to models and content generators evolved to evade stylistic detectors, and soon neither stylistic detectors nor expert humans will be able to tell the difference.

So platform filters will come to rely primarily on lab-encoded watermarks. If it’s economically valuable, labs will probably charge to detect the content their models generate (and to linit adversarial detection). Watermarks will themselves evolve to avoid removal.

Of course, only inference providers with an economic incentive or regulatory imperative will watermark the outputs they generate. Those without, ie. open source, user controlled, won’t. So “unaligned” (whatever that happens to mean) models and systems will come to outcompete human-generated content and “aligned” synthetic content along both volume and memetic axes.

So platforms, and governments, will be left with a choice: a) increase the cost of distributing all synthetic content (ie. by verifying account-level humanity and attaching real costs and consequences like member costs, hell bans, and legal consequences) to the point where it becomes unprofitable… best case is an email / spam equilibrium worst case a more dystopian ender’s game type lockdown; b) accept that models and systems with a provenance that is “unaligned” now own and drive the narrative across these platforms, and accept whatever slants and biases their creators bake i to those systems; c) or remove the filters and hope that a national hero model can outcompete the field.

My bet is on the first and a coin flip between the two sub-outcomes on a country-by-country basis, though I don’t think the last is out of the question.

[Dario Amodei @DarioAmodei]

2/2 Second, on the messaging around AI.  I do not agree that my messaging has been disproportionately negative.  In fact it has been about equally balanced between risks and benefits: I’ve written one major essay about each, and even in interviews where I discuss the risks, I make sure to frequently mention the incredible benefits as well as proposing possible solutions to the risks (short clips from my interviews that end up on social media tend to be disproportionately negative, as that gets clicks).  In fact, I wrote Machines of Loving Grace because I didn’t feel the AI industry was painting an inspiring enough picture of how the technology could radically transform the world for the better.  The bulk of the essay is devoted to refuting skepticism of AI’s potential in health and biology, and showing why I think it will actually be possible to cure most human disease in ~5-10 years, as crazy as it may sound to ordinary people and frankly to biologists as well (I used to be one!).  And, if you read my most recent essay (Policy on the AI Exponential), I discuss concrete proposals for how to streamline the FDA process to make sure the deluge of AI-accelerated drugs isn’t slowed down by the regulatory process.  I feel the urgency here: I lost my father to Hepatitis C only a few years before the development of direct-acting antivirals (sofosbuvir), which cure 95% of patients and probably would have cured him.

I do agree that the public has a negative view of AI (and that this is a big problem), but I don’t think it is primarily caused by me or any other AI leader warning about AI’s risks.  I think it is fundamentally a crisis of trust.  I think that ordinary people don’t trust companies, governments, or the tech industry and always suspect that we are cooking up some new way to screw them over.  The causes of this go back decades and AI is just the latest iteration of it.  I don’t think that a glitzy marketing campaign with a positive spin (which some have advocated that Anthropic do) is the way to win back that trust — at this point, saying that AI will cure cancer is more a cliche than it is inspiring, and most people think it is deceptive.  The thing that will work is *actually curing cancer*.  I think by far the most accurate criticism of AI companies including Anthropic is that we haven’t yet delivered on our big promises to benefit the world.  That is totally on us, and I think it’s the criticism you should be making, instead of all this stuff about messaging and marketing.

We are however doing our best to fix this: Anthropic is ramping up its efforts very quickly in biology and medicine, and we hope to have incredible results in the coming years and some early glimmers in the coming months.  When we’ve actually accomplished something real, the whole world will hear about it, as loudly as possible, you have my word on that.  But until then I don’t want to make empty promises, and in the meantime I feel compelled to speak honestly about the very real risks of AI and how to address them.  Honesty is the right thing on the merits, and in terms of public credibility and trust it is no worse than, and may in fact be better than, an approach that ignores or distracts from risks which people instinctively understand are real.

[Michael F @michaelfives]

Wish I could buy the dip on Anthropic rn

[Michael F @michaelfives]

@AmandaAskell The weird thing is also that the definition of harm often tends to center around cost. “Frequent scans lead to more frequent expensive follow up tests” misattributes the source of a systemic problem. Indict the cost of the follow up, not the scan.

[Michael F @michaelfives]

Seems inevitable that the outcome of this brouhaha is that the big labs will sandbag new model releases.

Which means if you are top talent and want to work on the frontier, the incentive to work with the big 3 is stronger than ever.

Feels like a loss for OS and innovation.

[Michael F @michaelfives]

Co-sign.

Surrogate endpoints and expedited IND’s probably matter most here … BUT given the arc of bio x ai, an FDA evidence library for reusable safety, validation, and translational data would tilt the world.

Would love to see something like that on this list.

[Joe Lonsdale @JTLonsdale]

1/ US biotech is in crisis, right before AI should be saving millions.

China is stealing away our industry and has surpassed the US in blockbuster pharma deals.

The next FDA Commissioner must be a fighter, and have a plan to overhaul the agency, beat China, and unleash cures.