For two decades, researchers debated whether adeno-associated viruses, the small viruses used to deliver new genes into thousands of patients, could in certain cases drive tumors. That debate is over.

A series of cases this year has presented powerful evidence that adeno-associated viruses, or AAVs, can cause cancer, forcing researchers and families to weigh the newly confirmed risk against the technology’s life-saving potential — and hopefully determine how to design better therapies in the future.

“It will change the calculus for some patients,” said Sam Mackenzie, a neurologist at University of Rochester, speaking specifically about spinal muscular atrophy patients. 

Continue to STAT+ to read the full story…

In Eating The Seed Corn I wrote about AI's destruction of entry-level jobs. But Karen Weise's Microsoft Offers Buyouts to 60,000 Employees as It Spends on A.I. reports on a different impact (my emphasis):

Microsoft on Monday offered buyouts to a quarter of its work force, the latest effort by the company to rein in costs as it spends heavily on artificial intelligence.
...
The “time-limited, voluntary” program grew out of the company’s first ever buyouts earlier this year, she wrote. In April, Microsoft offered early retirement to 8,600 long-tenured U.S. employees, and about 30 percent accepted the packages and left over the summer. Microsoft said it heard from additional employees who would have taken the buyout if they had been eligible. That informed the company’s decision to create this new, expanded program.

Microsoft expects many fewer acceptances in this new program, but the people leaving are not entry-level, they are well-paid senior staff. If the goal is to divert cash from salaries to capex getting rid of senior staff is much more effective.

Below the fold I look into the non-obvious implications of AI's impact on well-paid employees' jobs.

The buyout program isn't the only predator stalking Microsoft's employees:

Microsoft has already slightly shrunk its work force even as business has boomed. For several years, the company maintained a flat head count of about 228,000 employees.

But as of June 30, the end of its fiscal year, Microsoft’s work force had edged lower to 223,000 employees. Over that same fiscal year, Microsoft brought in $332 billion in revenue, up 18 percent from the previous year, and profit grew 31 percent to $134 billion.

In July, Microsoft announced it was cutting 4,800 jobs, particularly in its video game division. The company is scheduled to report quarterly results on Wednesday.

Many of these lost jobs will have been lower on Microsoft's pay scale, but still well-paid relative to the US average. The victims will have lost not only their salary, but also after a delay their health insurance. Crucially for this post, their and Microcoft's contributions to their 401(k) pension will have stopped. In the current economy they will struggle to find a new job.

Is this just a problem for the victims? In The Retirement Machine: AI, high earners, and the hidden risk to global stock markets Duncan Lamont asks:

how consequential would it be for equities if the US experienced a rise in AI-linked unemployment among high earners which led to a fall in pension contributions.

This turns out to be a really interesting question. Lamont writes:

In their AI-boom scenario, our economists model unemployment rising to 10% as workers are displaced by technology. Importantly, whereas past technological revolutions have displaced lower earners, this time higher-earning white-collar workers are in the firing line. And they are the ones who disproportionately contribute to the monthly pension flows into equity markets. This article puts numbers on just how material a risk this is. Spoiler alert: it’s big.

Lamont Figure 1

US pensions are a seriously big business. Lamont's Figure 1 shows that the value of US pensions investments are 47% bigger than all non-US stock markets combined, ond 82% of the US stock market. Lamont points out that:

The (dollar-weighted) average allocation to equities across 401(k)s is 75%. Younger individuals have more, with those in their 20s having 90% on average but even those in their 60s still hold over 60%.

How big are the "monthly pension flows into equity markets"? Lamont looks at 2024:

Assuming the 75% equity allocation also applies to new contributions, US DC plans and IRAs funnelled around $735 billion into global equity markets in 2024. Given the US weight in the global stock market and typical [target duration fund] allocations, 60-70% of that will have gone into US equities, translating into approximately $480 billion. DC and IRA pension accounts bought an estimated nearly half a trillion dollars of US equities in 2024.

DC (Defined Contribution) plans are 401(k) and similar plans to which employees and (typically) companies contribute monthly. Lamont estiimates that their purchases represent about 26% of the net purchases of US equities that year. Because this is a steady, predictable demand it plays a major role in keeping the US stock market growing in good times and cushions the fall in bad times.

Lamont Figure 3

Which employees contribute to DC plans? Lamont's Figure 3 shows that participation is much higher in industries that pay well. So there aren't just more of them, but they contribute more each month. The 26% keeping the US market afloat depends heavily upon the contributions of higher-paid employees.

Lamont cites evidence to support this skew. First:

The Congressional Budget Office’s most recent comprehensive distributional analysis of retirement-related tax expenditures — the closest available proxy for the distribution of contribution dollars — found the highest-income quintile captured 63% of the tax benefits in 2019, against 1% for the bottom quintile. The tax-benefit share will overstate the contribution share to some extent because higher earners face higher marginal tax rates. Even so, this is broadly consistent with IRS data which shows that 62% of contributions to IRAs in 2023 came from people earning over $100k a year. Only about 15% came from people earning what was then the median income of around $50k or less.

Lamont Figure 4

And second:

There is also direct evidence from within company plans. As Figure 4 shows, the best paid 20% of employees at a given firm have, on average, taken in 44% of employer contributions.

It is these these top earners who are at risk:

A growing body of research reaches the same broad conclusion: higher-income, white collar occupations requiring postsecondary education show the highest exposure to AI capabilities (Brynjolfsson and others (2018); Webb (2020); Muro and others (2019); Felten and others (2023); Eloundou and others (2024); Kinder and others (2024); and Hampole and others (2025)).

Dodini & Smith Chart 3

As I disucussed in Eating The Seed Corn, Stanford reseach concludes:

1. We find no evidence of widespread, economy-wide job displacement.
2. However, employment of young workers (ages 22–25) in AI-exposed occupations now stands 19% below where it would be had it kept pace with that of their less-exposed peers; experienced workers show no comparable gap.
3. This divergence has widened steadily since we first documented it in August 2025.
4. It operates primarily through reduced hiring of young workers rather than increased separations.

Lamont points out that this isn't good news:

You could take the junior-first pattern that the data currently shows as reassurance for the pension channel since junior workers contribute less than managing directors. But that would be naive. Junior workers in high-paying industries are still high earners by economy-wide standards. A 25-year-old software engineer on $140,000, or a first-year banking analyst on $120,000, sits comfortably in the top quintile of all US earners. The displacement happening right now is concentrated in industries where even the juniors out-earn the median American worker by a factor of two. These are exactly the contributors identified earlier. In addition, participants in their 20s hold 90% equity on average, the highest allocation of any age cohort. So even a junior-first shock would hit a group whose pension contributions are unusually equity-heavy.

Just as importantly, junior-first describes where current AI tools are biting first. We don’t know where it will end. The substitution frontier has moved up the capability curve every year since 2022, and there is no obvious reason to expect it to stop at the entry level.

After staring at the barrage of AI adds on the London Tube in summer 2025, I wrote The Selling Of AI:

The irresistible pitch to CEOs is that they can "do more with less", or in other words they can lay off all these troublesome employees without impacting their products and sales.

The effect is that CEOs will fire the workers first based on the ads and the hype, before finding out whether the AI agents can really replace them, and when they find out they can't will be unwilling, and unable to afford, to hire them back. And if they hype is right the upward movement of the "substitution frontier" will have the same effect — hugher unemployment for longer.

Either way the result will be a reduction of the flow of pension funds into the stock market. Lamont estimates:

A 10% hit to that exposed flow would remove roughly $20–30 billion of annual equity buying; a 25% hit would remove $50–75 billion; a 50% hit would remove $100–150 billion. In a normal year that is material. In a downturn, when buybacks, foreign demand and retail risk appetite are likely to weaken at the same time, it could matter a lot more.

As I write this the financial press is flooded with headlines such as:

• Stocks Slump as Wall Street Gets AI Wake-Up Call
• Amazon, Alphabet, SpaceX, and Meta are all to blame for the new tech stock rout
• Tech Stocks Tumble, but Broader Market Is Holding On
• Nasdaq 100 Hits Third Correction of 2026 on Unwinding Tech Stock Bets
• Meta, Alphabet, Intel Share Sales Leave Buyers Far Underwater
• US Tech Stocks Enter Correction on AI and Midterm Election Risks

The downturn may have already started.

The Peter Principle posits that people are promoted until they reach their level of incompetence. What happens after that point? The implicit assumption is that they stop getting promoted and linger at whatever level they’ve arrived at, and surely all of us can point to mediocre leaders who seem to exhibit exactly that circumstance. But in fact, that situation is not stable but subject to career pressure, as Christian Gläßel and Adam Scharpf define it in Making a Career in Dictatorship. Unable to move up, and very possibly unable to meet expectations, the person who is subject to the Peter Principle is at risk of losing their job, or even their entire career. Which means they might take an even greater risk to save it. Gläßel and Scharpf write:

The logic of career pressure also offers a compelling twist to the Peter Principle. The Principle posits that individuals in hierarchical organizations are promoted until they reach their level of incompetence, where they can no longer perform effectively and cease to advance. In contrast, our logic shows how these individuals may surpass this point and climb the hierarchy even further. Career-pressured individuals often take on high-risk, labor-intensive, or morally contentious tasks. By loyally executing the organization’s dirty work, they can reestablish their value, regain competitiveness, and even outpace peers who were previously ahead of them. Alternatively, they may resort to sabotaging superiors or higher-ups by making false allegations or revealing secrets to create vacancies above them and forcibly advance their careers.

Gläßel and Scharpf, Making a Career in Dictatorship, page 182

In other words, someone promoted up to their level of incompetence, sensing a threat to their own livelihood, becomes a threat to everyone else. I’m reminded here of David Graeber’s theory in Bullshit Jobs, that a great many management jobs are basically feudal retainers—people kept around in order to make the folks above them look good. Very often, making someone else look good may well mean tackling the dirty work for them so that they can keep their hands clean.

The petered leader has in effect been made an imposter—a real imposter, not a “syndromed” or perceived one. They must seem competent, to preserve the appearance that they belong in the role; while in fact they are not capable of delivering on their actual responsibilities and may have been quietly tasked with unrelated services both illegal and unethical. As a result, one of the primary tools of the real imposter is obscurity. Imposters cloak themselves in a fog of obfuscation and busyness, so that the people around them can never say for sure what it is they are doing, or how well (or not) they are doing it.

What we call imposter syndrome was originally termed imposter phenomena, a term that makes plain that it is something that can be observed, not merely felt. The transformation into a syndrome shifted the experience of being impostered by a patriarchal, white supremacist society into a medicalized disorder, something for which the impostered person must be treated. That’s a neat trick: to shift the locus of treatment to the mistreated person, rather than the system that caused her harm. But perhaps there’s even another, more diabolical move here. What if the popularity of imposter syndrome is, in part, a means of obscuring the actual imposters in our midst? What are we missing when we look inward instead of around?

---

View this post on the web, reply via email, or become a supporter.

The best reality TV show I’ve ever seen*–indeed one of the best TV shows I’ve ever seen, period**–is the original British version of Faking It. It ran on Channel 4 in the early 2000s, when I was a postdoc in the UK. A member of the public would take one month to learn to fake it as an professional in a field mostly or entirely unrelated to their real profession. They would learn to do this by moving in with one or two mentors–themselves experienced professionals in the field in which their mentee had to fake it–and doing intensive training full-time. The climax of the show was a contest pitting the faker against three actual professionals, each with a few years of full-time professional experience. The contest was judged by elite pros in the field. The judges would decide who performed the best. Then they’d be told that one of the contestants was faking it–had only been doing it for a month–and asked to guess who it was. The faker won if they won the contest and weren’t ID’d as a faker by any of the judges. There wasn’t any prize for winning, per se, but the people who took part in the show invariably got a lot out of it (more on this below). You can find old episodes on YouTube, and you really should do yourself a favor and check them out. Some of the best ones include cellist to club DJ, burger flipper to chef, and physics grad student to magician.

Faking It was an amazing show. It changed the lives of a bunch of the people who went on the show. Remember, this was pre-influencer times–the contestants weren’t looking to become influencers or reality TV stars. They weren’t even looking to change their lives, at least not necessarily. But yet they did. Some people who went on this show changed professions. They literally faked it ’til they made it! And even those who went back to their old professions and old lives often went back as completely different people–much more self-confident, much more outgoing, and with totally different interests, to the point that their old friends and colleagues were shocked at the changes. I love the message that you don’t just have to “be yourself,” not even “the best version of yourself.” You can be different, and better, than that. People can really change, if they really want to, and if they’re placed in the right environment.

Which got me wondering if Faking It has any lessons for mentoring more broadly, in particular for academic scientists mentoring graduate students. Obviously, Faking It is very different than grad school in all sorts of ways–but that’s the point. You don’t just have to look close to home for insights and inspiration. Looking very far afield can help too.

Mentoring lessons from Faking It

“Faking it” is a misnomer. The phrase ‘fake it ’til you make it’ is easily misinterpreted, because the ordinary connotation of ‘fake’ is something with no basis in reality whatsoever: an actor playing a part, or an AI-generated video clip, or one of Danielle Dixson’s datasets. But the mentees on Faking It really did master many aspects of their new profession. Now, they generally did so in circumscribed ways. For instance (mild spoiler alert), IIRC the sheep shearer who became a hairdresser only learned one haircut, because he only needed to learn one for the contest at the end. I do think there’s a lesson here for many graduate students (and postdocs and junior faculty) in ecology, especially those suffering from imposter syndrome. You’re not an imposter–not a fake with no basis in reality. Rather, you’ve already mastered some aspects of your profession, and you’re in the process of mastering more.

There’s talent everywhere. Previous experience is overrated. Faking It mentees generally had little-to-no relevant previous experience in their new professions (although see below for exceptions). But yet they almost all faked it successfully or came quite close.

Desire, determination, and grit are properly rated. Faking It mentees knew what they were getting into, at least in broad strokes. They knew they were signing up for a very intense month of training, very different from anything they’d done before. And they saw it through. They were a self-selected bunch of gritty, determined people, though the environment they were in also fostered their grittiness and determination (more on that below).

Technical skills are both the easiest and least-important things to learn. A common theme of Faking It episodes is that learning technical skills isn’t the hardest part, or the most important part, of learning a new profession. Apparently, anybody who really wants to can learn the technical skills required to perform card tricks, or mix tracks as a DJ, or filet a salmon, or whatever. Over and over again, the hard part of faking it almost always turns out to be acquiring the personality traits and soft skills you need to fake it in your new profession. You invariably need self-confidence. And depending on the profession in which you’re faking it, you need other personality traits and soft skills too. For instance, to be a magician you have get comfortable with lying to people. To be a chef, you have to learn to be calm, focused, and situationally aware while working as part of a team under intense deadline pressure. To be a club DJ, you have to be outgoing and enjoy partying with other people. Etc.

Tutoring beats letting people figure things out for themselves. Grad students often are expected to figure a lot of things out for themselves and to teach themselves a lot of things. Many graduate advisors, including me, take the view that this is how you learn to think and work independently–you learn it by doing it. But that’s the polar opposite of how mentoring works on Faking It. Mentors on Faking It spend a lot of time tutoring their mentees one-on-one, and in many cases that seems to have set the mentees up for success as independent professionals in a new profession. As a mentor, I wonder if I have room to improve in recognizing when one of my graduate students needs me to sit down for a long time with them, teach them something, and give them drills/assignments/etc. to help them master whatever it is I’m teaching them.

Figuratively throwing people in at the deep end is a great way to teach them to swim. Or at least, it’s a great way to start their first swimming lesson. Mentors on Faking It routinely threw their mentees in at the deep end from very early on–surprising them by asking them to fake it in their new profession for members of the public within their first few days (sometimes even on their first day!). No doubt that was to create some drama for the TV show. The TV audience got to watch through their fingers as the mentees invariably crashed and burned at their first attempt at faking it. But just because it made for good TV, and just because it invariably made the mentees incredibly stressed and embarrassed, doesn’t necessarily mean it was bad mentoring–that it hindered rather than aided the mentees at achieving their ultimate goal of faking it successfully. Throwing the mentees in at the deep end invariably gave the mentors insights as to the mentees’ strengths and weaknesses. It also created a teachable moment, opening the mentees’ eyes to just how much work it was going to take for them to fake it successfully. And it motivated the mentees, a point that deserves its own bullet…

Failure, and risk of failure, can be great motivators. The mentees on Faking It get subjected to a series of public tests throughout their training, in preparation for the final contest. They almost invariably find those tests nerve-wracking, and they sometimes fail those tests quite badly. Almost everybody hates failure, and especially hates embarrassing, public failure. The mentees on Faking It invariably channel that fear of embarrassing failure into intense practice. To be clear, the mentors on Faking It aren’t cruel–they’re not deliberately setting their mentees up for failure purely for the sake of having them fail. Rather, they’re pushing their mentees and setting high expectations, recognizing that if mentees never fail during their training it means the expectations weren’t high enough, and that failing to push the mentees would merely set them up for a huge failure in the final contest. This broadly jives with my own experience as a mentor. Several of the cases in which my own mentees made the greatest strides were cases in which they were very anxious about an upcoming high-stakes task at which there was some real (or at least perceived) risk of failure: an oral candidacy exam, an oral thesis defense, a presentation at a big international conference. And just like on Faking It, my own mentees invariably gain a ton of confidence from succeeding in a situation in which there was some real, or at least perceived, risk of failure. On the other hand, few of my own graduate students have experienced actual failure during their time in my lab. More broadly, I have the (admittedly anecdotal) sense that few grad students anywhere ever fail their candidacy exams or thesis defenses these days. I wonder a little if that means I, and others, need to adjust our mentoring approaches, whether by setting higher expectations, or by arranging for student to have more opportunities to engage in tasks on which failure is possible.

Fear of letting others (and yourself) down can be a great motivator. Mentors and mentees on Faking It often do short talks to the camera about how they think the training is going and how they’re feeling about it. A common theme is that mentees don’t want to let themselves down. They want to feel like they’ve given their best effort, want to be able to look themselves in the mirror and feel proud of themselves, and don’t want to feel that they’ve wasted their time. But mentees aren’t focused solely on themselves–they also don’t want to let their mentors down. They invariably like and trust their mentors, and deeply appreciate all the time and effort (and, often, professional and personal sacrifices) that their mentors have put into training them. All mentees on Faking It feel like they owe it to their mentors to succeed. And for their part, the mentors invariably like the mentees, and also feel that the mentees owe it to both themselves and their mentors to put in their best effort. This is a point I want to reflect on more. Over the years, I’ve increasingly emphasized to my graduate students that their graduate program is for them, not for me–that they don’t work for me and don’t owe me anything, and that it’s my job to help them achieve their goals. Have I swung too far in that direction? Or perhaps this is an area in which the concentrated intensiveness of Faking It makes it disanalogous to grad school? The mentors on Faking It spend a lot of time training (and living with!) their mentees over the course of a month. If someone volunteers to put you up for a month and spend all day every day training you, you might rightly feel that you owe them more than you owe a graduate supervisor who meets with you once a week for an hour or something.

Good mentoring focuses on the things the mentee most needs to improve, and plays to the mentee’s strengths. The mentors on Faking It focus their training exercises on the things the mentees most need to improve at. “Face your fears” seems to be the motto of every Faking It mentor. No mentee can ever get out of a training session or training exercise by saying–truthfully!–that it makes them feel anxious or uncomfortable. But Faking It wasn’t solely about mentees improving their weak points and facing their fears; the mentors also helped the mentees play to their own strengths and work around their weaknesses as far as possible. For instance (very mild spoiler incoming), they helped the physics student-turned-magician to develop a stage show that wouldn’t require him to talk much when he first came out on stage, because he was struggling to master stage patter. In summary: when deciding whether to focus on improving a mentee’s weaknesses, or playing to a mentee’s strengths, Faking It mentors answer “Why not both?”

It’s great to change your mind–but not too often or too easily. An implicit premise of Faking It is that on the one hand people have a lot of agency and a lot of options in life–more of both than they probably realize. You can choose to be something, and someone, totally different than what you are right now. But that on the other hand, you can’t exercise your agency, and choose among your options effectively, if you change your mind too often or too easily. Once you’ve decided to go all-in on faking it as a chef or a magician or whatever, you have to follow through; you can’t back out at the first (or even second, or third) hurdle. Faking It suggests that you shouldn’t be either too slow or too quick to change your mind about what you want to be.

It’s sometimes surprising what previous experience is or isn’t ‘relevant’. Some mentees had a bit of previous experience that you’d have thought would be helpful, but turned out not to be helpful at all. Conversely, a few mentees’ previous professions turned out to be shockingly relevant to their new professions.*** (SPOILER IN FOOTNOTE) I’d say my own experience as a mentor for grad students has confirmed this. Some of the best grad students I’ve encountered were excellent in part thanks to previous experience that doesn’t have any obvious direct relevance to grad school in ecology (e.g., a background in humanities; seemingly-unrelated private sector work experience).

Some things can’t be faked, or rushed. The least successful episode of Faking It that I saw involved teaching a guy with a desk job to fake it as a pro surfer. The guy failed, badly–he couldn’t pass as a pro surfer physically, nor could he do anything beyond riding a surfboard straight to the beach. It’s simply not possible to get physically fit enough in a month to pass as a pro surfer, unless you’re pretty fit already (which this guy wasn’t). Analogously, I’m sure there are some aspects of being a good a scientist that you can’t learn–or even learn to fake–in a short amount of time.

What do you think? As I hope is obvious, Faking It does not provide a complete, exhaustive Theory of Mentoring, and the above is not a complete, exhaustive List of Things I Believe About Grad Student Mentoring. I just found it interesting to mull over Faking It and reflect on whether, despite its obvious differences from grad student mentoring, it might nevertheless provide some transferable lessons. Looking forward to your comments as always. I encourage you to use the blog’s comment thread, but you can also use the anonymous comment form at the bottom of the post.

*Small sample size alert

**Small sample size alert #2

***It turns out that being a professional cellist is excellent preparation for being a club DJ. Or, perhaps, being a talented musician is excellent preparation for being both a professional cellist and a club DJ.

Cop30 host sets out steps to break link with coal, oil and gas as it aims to move focus from negotiating to implementing

All countries can start to break their dependence on fossil fuels now without needing to wait for new technologies or economic changes, Brazil’s government has said ahead of a global climate summit.

Brazil, the host of last year’s UN climate conference, set out the world’s first global “roadmap” for phasing out fossil fuels on Tuesday, showing that countries could boost their economies and pull people out of poverty while breaking the link with coal, oil and gas.

Continue reading...