Happy New Year friends, we’re back with good news for the balding, bad news for emulsifiers and another step towards uploading yourself to the cloud in the event the new Alzheimer's breakthrough flops. Enjoy!

Only a short while after fatness turned optional with GLP-1s, baldness may be going the same way. So long to the fat bald man.

Cosmo Pharmaceuticals has just posted Phase III results showing 539% relative improvement in hair growth. There have been other drugs on the market for years, notably finasteride, but they've come with the sort of trade-offs that made many men decide a receding hairline was the lesser evil.

Finasteride works systemically, blocking hormone production throughout your entire body, which occasionally results in side effects nasty enough to make men pull their new luscious hair out. Clascoterone 5% takes a different approach, blocking dihydrotestosterone directly at the hair follicle, meaning your hormones remain unmolested everywhere else.  If all goes according to plan, this will become the first topical androgen receptor inhibitor for hair loss and the first new mechanism in over three decades.

🧐 What's in it for me? Get that ruler off your forehead, relief could arrive within 18 months. The 12 month safety follow-up wraps in spring 2026, regulatory submissions follow, and then we enter the thrilling world of pharmaceutical approval timelines. Clascoterone already exists as a 1% acne cream, which means the compound isn't entirely unknown to regulators and may ease the process.

💵 Out of the Lab: Male pattern baldness affects over a billion men globally, but the market's been stuck with two molecules since the '90s, one that works systemically with unfortunate side effects, one that works topically with variable results. Clascoterone could be the first genuine innovation in a generation.

Brain-computer interfaces have traditionally required surgeons to remove sections of skull to accommodate electronics roughly the size of a deck of cards, which unsurprisingly, has been quite a tough sell. Now though, a new system (BISC) fits on a chip so thin you could slip it between skull and brain, no saws required.

The numbers are impressive: 65,536 electrodes, 100 Mbps wireless data transfer and a footprint of 3 cubic millimetres. For context, that's faster than home broadband and smaller than a grain of rice, though considerably more expensive. 

The practical applications run from the conservative (epilepsy management, restoring motor function) to the terrifying (brain-to-AI interfaces), though the researchers are diplomatically focusing on the medical uses for now. Short-term human studies are now underway.

🧐 What's in it for me? For now, the immediate beneficiaries (i.e. within 10 years) are people with neurological conditions who've been waiting for implants that don't require the surgical equivalent of home renovation.

💵 Out of the Lab: Brain-computer interfaces are transitioning from lab curiosities to commercial products, but most designs still require bulky hardware or penetrating electrodes that damage cortical tissue. If BISC's chip continues to perform through trials, we may see a total paradigm shift. 

Alzheimer's has been considered a one-way journey since 1906 and the standard approach has been to slow progression or, if you're feeling ambitious, delay onset. This makes a new study particularly interesting as they've gone and reversed advanced Alzheimer's in mice, restoring full cognitive function even after significant brain damage.

The trick involves NAD+, a molecule that keeps cells energised. It declines with age in everyone, but in Alzheimer's brains it collapses entirely. When researchers used a compound called P7C3-A20 to restore NAD+ levels, mice with severe cognitive impairment recovered completely. Not "showed improvement" or "demonstrated slower decline," but actually got better, and blood biomarkers used to diagnose Alzheimer's returned to normal. The effect worked in two genetically different mouse models suggesting this isn't a fluke. 

But before you go bulk-buy NAD+ supplements, the researchers are sensibly cautious as flooding your system with NAD+ precursors has also been shown to promote cancer in animal models. Out of the frying pan, into the big cancery fire. 

🧐 What's in it for me? If this translates to humans, the entire Alzheimer's treatment paradigm would shift from managed decline to potential recovery, which would be rather good news for the 55 million people suffering. Clinical trials will take years, and mouse results have a habit of not transferring to humans, so temper expectations accordingly but the conceptual breakthrough is that researchers are even testing for reversal rather than just prevention.

💵 Out of the Lab: Alzheimer's is one of medicine's most lucrative unsolved problems, with failed trials littering the landscape and neurodegeneration biotechs trading at valuations that assume disappointment. If energy metabolism proves more important than amyloid plaques, the entire treatment approach pivots.

Chipmakers have been running into two inconvenient laws of physics. The miniaturisation wall, where transistors refuse to shrink much further without exhibiting quantum mechanical tantrums, and the memory wall, where processors can compute faster than data can be delivered to them.

A new chip from a smorgasbord of US Universities and SkyWater Technology has solved this by building upward rather than outward, stacking computing and memory layers like an electric wedding cake.

Early results showed the 3D chip beating comparable 2D designs by about four times and simulations suggest twelve-fold improvements are possible as more layers are added. The chip uses monolithic 3D integration, meaning each layer is built directly atop the previous one rather than gluing separate chips together, which allows for vastly denser connections and avoids the bottlenecks that plague simpler approaches. The result is a chip that moves more data, faster, using less energy per operation, which is the semiconductor equivalent of having your cake and eating it.

🧐 What's in it for me? Frontier AI models are getting so absurdly large that hardware has become the limiting factor, which is irritating for anyone trying to run them. If 3D chips deliver, we can expect faster AI and breakthroughs in physics simulations that currently choke on data throughput. The immediate beneficiaries are hyperscalers building AI infrastructure and anyone trying to train models that don't fit in existing memory architectures.

💵 Out of the Lab: AI's exponential growth is hitting hardware limits and the memory wall is becoming the primary bottleneck for next-generation models. These 3D chips could be the architectural shift that enables the next decade of AI scaling, but only if it can be manufactured economically at volume.

In case you need some New Year’s Resolution inspiration, this one may lack the inspirational heft of "run a marathon" but it might actually do far more to improve your (and your offspring’s) health. 

Researchers fed emulsifiers to pregnant and nursing mice, then watched their offspring develop disrupted gut bacteria, chronic inflammation, and higher obesity rates, despite never consuming the additives directly. Maternal dietary choices, it turns out, have consequences that extend beyond the immediate inconvenience of pregnancy, and emulsifiers don’t seem to help anyone. 

They’re everywhere: dairy products, baked goods, ice cream and baby formula. They improve texture and shelf life, which manufacturers consider important and which scientists are now discovering, like many processed food ingredients, might not be entirely benign. The offspring developed disrupted gut bacteria that triggered chronic inflammation, despite never eating the additives themselves. Their immune systems skipped the training phase where they learn tolerance, leaving them with hair-trigger responses to everything.

Infant formula is particularly concerning, since emulsifiers appear when babies' gut bacteria are first establishing. The researchers are calling for stricter regulation and human trials, though regulatory agencies move with the urgency of continental drift. If you're pregnant or feeding an infant, check ingredient lists for carboxymethyl cellulose (E466) or polysorbate 80 (E433).

So there's your resolution sorted. Fewer additives, better gut bacteria, less inflammation. Considerably less exciting than finally joining that gym, but on the plus side, no lycra required. 

Until next time, stay curious.

Like what you're reading? Share toast with a friend.