The allure to report anything as a medical breakthrough is strong. So strong that even the Financial Times can’t avoid it:

New diagnostic technology that uses fibre optics to find the causes of heart disease has begun ⊕ Emphasis mine. clinical testing at London’s St Bartholomew’s Hospital.

The iKOr device, developed at Barts Health and University College London, measures blood flow around the heart. Researchers say it could eventually help many thousands of patients suffering from cardiovascular symptoms such as chest pains, whose cause cannot be identified with current techniques.

“This new device is a game-changer in how we manage heart disease, making it a lot easier to assess the health of a person’s heart,” said Anthony Mathur, clinical director for interventional cardiology at Barts.

Three patients have undergone testing to date, out of 10 planned in the first phase. Another 100 may, subject to regulatory approval, before the device could potentially become commercially available, if it’s demonstrated to work. There is, it goes without saying, no clinical data published to date.

How does this change the game, exactly, when we don’t yet know if it works? The use of undeserved superlatives in cancer drug reporting is well documented so it’s not a surprise to see cardiology, that other lucrative medical subspecialty, being much the same.

What is a surprise is seeing the usually reliable FT falling down to the level of The New York Times in spreding medical jingoism. How interesting that in both cases it was a local hospital — Memorial Sloan Kettering for NYT, St Bart’s for FT — serving as the source. So interesting that I have to think there were some personal behind-the-scenes goings on.

Yes, yes, America has terrible health care — even a tech podcast says so — while paying an order of magnitude more for it than other rich countries. But hear me out: what if the costs are so high because Americans are (unsuccessfully) trying to buy their way out of poor policy decisions, from dependence on cars, to the early 2000s' promotion of opioids, to the widespread availability of cheap but nutrient-less calories, and no amount of fiddling with who pays for what in healthcare will be able to fix that?

Which is to say: it’s fine to look at specific costs and specific outcomes — I have done so myself — but what exactly is the action item after reading a report like The Commonwealth Fund’s cited by Ars Technica?

Warning, it’s a press release:

Gaithersburg, MD, January 31, 2023 – Cartesian Therapeutics, a fully integrated, clinical-stage biotechnology company pioneering RNA cell therapies for autoimmune diseases and cancer, has dosed the first participant in its Phase 2b randomized controlled trial (RCT) for generalized myasthenia gravis (MG), an autoimmune disorder that causes muscle weakness and fatigue. The RCT will evaluate the efficacy and safety of the company’s lead asset, Descartes-08, a first-in-class, RNA-engineered chimeric antigen receptor T-cell therapy (rCAR-T).

To the company’s knowledge, this is the first placebo-controlled study of an engineered cell therapy, and the most advanced investigational cell therapy in clinical development for any autoimmune disease. Descartes-08 is administered over 6 weekly outpatient visits and requires no preconditioning chemotherapy.

The manuscript from the open-label study is almost done, but some of the data was presented back in September 2022 (and available on YouTube). I also talked about the study in an MGFA webinar.

RCTs FTW.

So let me see if I have this straight:

• OpenAI and Microsoft have a partnership, with Microsoft all-in on integrating aritifical intelligence into its products and developing it further.
• Google/Alphabet has its own AI programs — some of them mind-blowing — which are bound to increase now that ChatGPT is out with potential to destroy the value of the one thing Google still does well: searching the internet.
• Facebook/Meta has only some embarassing failures to show for their efforts, for now, but one can expect some rearrangement of priorities once the company’s shareholders see what being all-in on AI has done to $MSFT.

If artificial general intelligence is possible, ⊕ For an explanation for how AI differs from AGI I recommend this short interview with David Deutsch. odds are that it will emerge in this decade. Determining whether that is good or bad I will leave as an excercise to the reader.

P.S. While getting the links for this post I came upon a WaPo article which came out today and devotes a single paragraph to the potential harms of AI:

Some AI ethicists fear that Big Tech’s rush to market could expose billions of people to potential harms — such as sharing inaccurate information, generating fake photos or giving students the ability to cheat on school tests — before trust and safety experts have been able to study the risks. Others in the field share OpenAI’s philosophy that releasing the tools to the public, often nominally in a “beta” phase after mitigating some predictable risks, is the only way to assess real world harms.

This is true, as things stand now. Wouldn’t it be nice if it stayed that way.

Finding an article about AI in a major news publication that sticks to facts and makes sense has become an event worth celebrating, so here is a recent one by Tatum Hunter of the Washington Post.

Feeling sad for Twitter app developers and, considering there will be some delay between the fit hitting the shan and it spreading all around, even sadder for the inevitable hardship of anyone who depended on a Twitter audience for their livelihood. Castles out of sand…

Further evidence that T cells are the best cells.

The headline: ChatGPT appears to pass medical school exams, educators rethinking assessments.

The article:

• They were mock, abbreviated exams,
• done incorrectly, ⊕ There are no open-ended questions on the real USMLE.
• which it didn’t actually pass,
• and which were reported in a pre-print. ⊕ Which isn’t a complete knock against the study per se, but even a glance at it shows that some questionable choices have been made regarding the scope — there were only 376 publicly available questions instead of more than a 1,000 on the real exams — and the methods used to ensure the publicly available questions hadn’t already been indexed by the ChatGPT training algorithm.

To be clear: this is my complaining about misleading headlines, not saying that predictive AI wouldn’t at some point be able to ace the USMLE, that point not being now, for reasons stated above. And let’s not even get into whether having a high USMLE score means anything other than the person achieving a high score being a good test-taker (it doesn’t).

And with that, the Twitter chapter of my life has closed.

May 2008 to January 2023. Not a bad run, considering.

Adam Mastroianni’s Experimental History newsletter has enabled paid subscriptions today, and if there is one science-oriented Substack worth paying for, it’s Adam’s. I’m sold.