Daniel sent us this one — he's asking about the history of vaccines. How were they invented, how many exist today, and the part I'm most curious about: is vaccine technology still actually evolving, or are we just tweaking flu shots and calling it innovation? There's a real tension in that question, because the mRNA moment made a lot of people wonder whether we'd just hit peak vaccine and everything else is maintenance mode.
The answer is no, not even close. The technology is in the middle of what I'd call a genuine platform shift. But to understand why, you have to see how many times it's already shifted. Most people think the story is: Jenner, smallpox, then a long gap, then polio, then we figured it out. That's not the story at all.
I think what you're really saying is — buckle up, I've got a timeline.
I have a timeline and I'm not apologizing for it.
Wouldn't dream of asking you to. Let's hear it.
The starting point is older than most people realize. Variolation — deliberately infecting someone with a mild case of smallpox to protect them from the severe version — was practiced in China and India by at least the fifteen hundreds, possibly earlier. The Chinese method was to take dried smallpox scabs, grind them into powder, and blow them up the nose. Not exactly a pleasant Saturday afternoon.
The original nasal spray.
And it worked, sort of. Mortality from variolation was about two to three percent. Which sounds horrifying — unless you compare it to natural smallpox, which killed thirty percent of the people it infected. So in context, a two percent gamble was rational. Lady Mary Wortley Montagu, the wife of the British ambassador to the Ottoman Empire, observed variolation in Constantinople in seventeen seventeen. She had her own children variolated, brought the practice back to England, and basically became the eighteenth-century equivalent of a vaccine influencer.
A British aristocrat watching Ottoman medical practice and importing it to London. That's a nice corrective to the usual narrative of medicine flowing one direction.
And it set the stage for Edward Jenner in seventeen ninety-six. Jenner noticed that milkmaids who'd caught cowpox — a mild disease — seemed immune to smallpox. He took material from a cowpox sore on a milkmaid named Sarah Nelmes and inoculated an eight-year-old boy, James Phipps. Boy got a mild fever, recovered, and later proved immune to smallpox. Jenner called it "vaccination" from the Latin vacca for cow. That's where the word comes from.
Eight-year-old boy. Can you imagine trying to run that clinical trial today?
The ethics review alone would take longer than the entire eighteenth century. But that was the first deliberate, documented use of a related virus to provoke cross-protection. It's a different principle from variolation. Variolation gave you the actual disease in a controlled dose. Vaccination gave you a different pathogen that your immune system mistook for the real thing.
Which is the core insight that everything else builds on. So Jenner publishes in seventeen ninety-six. Because I know there's a gap.
Almost a century. And this is the part most people miss. Jenner proved the concept, but nobody understood why it worked. Germ theory didn't exist. Viruses hadn't been discovered. The immune system was a black box. So for eighty-plus years, smallpox vaccination spread globally — it became the first vaccine actually eradicated a disease, though that wouldn't happen until nineteen eighty — but no new vaccines were developed. The tools for invention didn't exist because the science wasn't there.
The bottleneck wasn't imagination. It was biology.
Louis Pasteur broke the bottleneck in the eighteen eighties. He was working on chicken cholera and famously left a culture out over the summer. When he injected it into chickens, they got mildly sick and recovered. Then he injected them with fresh, virulent bacteria, and they survived. The accidental attenuation had weakened the pathogen enough to trigger immunity without causing severe disease. Pasteur called it a "vaccine" in Jenner's honor, even though there were no cows involved.
Which is a lovely bit of scientific etymology. Naming your discovery after someone else's cow.
Pasteur then developed vaccines for anthrax in eighteen eighty-one and rabies in eighteen eighty-five. Rabies is fascinating because it was the first vaccine made from a pathogen grown in an artificial medium — rabbit spinal cord tissue — and then deliberately weakened. It was also the first therapeutic vaccine. You could give it after exposure because rabies has a long incubation period. The immune response could outrun the infection.
By the end of the nineteenth century, we've got smallpox, anthrax, rabies. What's the count today?
The World Health Organization currently lists vaccines for more than twenty-five diseases. The exact number depends on how you count combination vaccines — the MMR is one shot but three diseases. But if you count distinct diseases we can prevent with vaccines, it's somewhere between twenty-five and thirty. Diphtheria, tetanus, pertussis, polio, measles, mumps, rubella, hepatitis B, hepatitis A, varicella, pneumococcal disease, meningococcal disease, rotavirus, human papillomavirus, influenza of course, yellow fever, Japanese encephalitis, typhoid, cholera, rabies, tuberculosis with BCG, and now COVID-19 among others.
Twenty-five to thirty diseases. And that's just the ones with widely approved vaccines. How many of those came in the first half of the twentieth century versus the second?
Very few in the first half. The big wave started in the nineteen forties and fifties, and it's tied directly to the invention of cell culture. Before cell culture, you needed a living organism to grow viruses — chicken eggs, animal brains, live animals. That was slow, expensive, and introduced contamination risks. John Enders, Thomas Weller, and Frederick Robbins figured out how to grow poliovirus in non-nervous tissue culture in nineteen forty-nine. Won the Nobel Prize for it. That single technique unlocked the polio vaccine — both Salk's inactivated vaccine in nineteen fifty-five and Sabin's oral vaccine in nineteen sixty-one — and also opened the door for measles, mumps, rubella, and varicella vaccines.
Cell culture was the first platform shift. What's the full taxonomy if you're mapping the evolution?
I'd say there are roughly five platform eras. First, live attenuated and inactivated whole-pathogen vaccines — that's Jenner through Pasteur through mid-twentieth century. You take the whole virus or bacterium and either kill it or weaken it. Second, subunit and toxoid vaccines — you don't use the whole pathogen, just a piece of it, like a surface protein, or an inactivated toxin the pathogen produces. Diphtheria and tetanus vaccines are toxoids. Hepatitis B vaccine, introduced in nineteen eighty-six, was the first subunit vaccine — it uses a yeast-derived surface antigen.
Hep B was also the first recombinant vaccine, right? Made by genetic engineering rather than growing the actual pathogen?
That's the third era — recombinant vaccines. Instead of purifying protein from the pathogen itself, you insert the gene for the antigen into yeast or bacteria and let them manufacture it. Much safer, much more scalable. Then the fourth era is conjugate vaccines. Some bacteria — like the ones that cause meningitis and pneumonia — have a polysaccharide capsule that hides them from the immune system, especially in young children. Conjugate vaccines chemically link that polysaccharide to a protein carrier that the immune system recognizes, which trains it to respond to the capsule. The Hib vaccine, introduced in the late nineteen eighties, is the classic example. Before Hib conjugate vaccines, Haemophilus influenzae type b was the leading cause of bacterial meningitis in children under five. Now it's virtually eliminated in countries with routine vaccination.
The fifth era is mRNA.
Viral vector vaccines. I'd bundle those together as the genetic platform era. Instead of delivering the antigen, you deliver the genetic instructions for the antigen — either as messenger RNA or packaged in a harmless virus — and let the body's own cells produce it. The COVID vaccines were the first widely deployed mRNA vaccines, but the technology had been in development for decades. Katalin Karikó and Drew Weissman published the key paper on nucleoside-modified mRNA in two thousand five. They figured out how to modify the mRNA so the immune system doesn't immediately destroy it before it can do its job.
The COVID mRNA vaccines weren't a bolt from the blue. They were the overnight success that took thirty years.
That's the pattern with vaccine technology. Each platform shift builds on a long period of incremental work, then a crisis or a breakthrough pushes it over the line. What's different now is that the pace is accelerating and the platforms are multiplying.
Let's stay on that acceleration. What's actually in the pipeline that's genuinely new, not just a new target for an existing platform?
Let me organize this. One major frontier is universal vaccines — vaccines that cover all variants of a pathogen so you don't need annual reformulation. A universal flu vaccine is the holy grail here. Current flu vaccines target the head of the hemagglutinin protein, which mutates rapidly. A universal vaccine would target the stalk, which is conserved across strains. There's also work on universal coronavirus vaccines that would protect against future pandemic coronaviruses, not just SARS-CoV-2.
The stalk versus the head. That's such a clean metaphor for what good vaccine design does — find the part the virus can't afford to change.
And that's only possible because of advances in structural biology. Cryo-electron microscopy lets us visualize viral proteins at near-atomic resolution. You can literally look at the structure of a protein and identify conserved regions that are functionally essential. That's a completely different approach from the old empirical method of "grow it, kill it, inject it, and see what happens.
The visual equivalent of aiming for the stalk.
Second frontier: therapeutic vaccines. Vaccines that treat existing disease rather than preventing infection. Cancer vaccines are the most prominent here. There are already approved therapeutic cancer vaccines — sipuleucel-T for prostate cancer was approved in twenty ten, though it's not widely used. But the newer approach uses mRNA to encode neoantigens — mutated proteins unique to a patient's tumor. You sequence the tumor, identify the mutations, design a custom mRNA vaccine that teaches the immune system to attack cells expressing those mutations, and manufacture it per patient.
That's personalized medicine at a level that would have been science fiction twenty years ago.
BioNTech and Moderna both have personalized cancer vaccines in phase two and phase three trials. Moderna's candidate, in combination with Merck's pembrolizumab, showed a forty-four percent reduction in risk of recurrence or death in high-risk melanoma patients in a phase two trial published in twenty twenty-four. The phase three trial is ongoing.
We're talking about a vaccine that is designed for one person, based on the genetic sequence of their tumor, manufactured on demand. That's not "tweaking flu shots.
Third frontier: needle-free delivery. Microneedle patches, intranasal vaccines, oral vaccines. An intranasal COVID vaccine was approved in India and China. The logic is that respiratory pathogens enter through the mucosa, so you want mucosal immunity — IgA antibodies right at the point of entry — rather than just systemic immunity from an intramuscular injection. It's like having security at the door instead of just in the building's interior.
Which also eliminates needles, cold chain requirements potentially, and the need for trained healthcare workers to administer shots. That's a distribution revolution, not just an immunological one.
It might reduce transmission. Injected vaccines are good at preventing severe disease but less effective at preventing infection and transmission in the upper respiratory tract. Mucosal vaccines could close that gap. Fourth frontier: self-amplifying RNA. Traditional mRNA vaccines deliver a fixed amount of mRNA that gets translated and then degrades. Self-amplifying RNA includes sequences — derived from alphaviruses — that encode an RNA replicase. The RNA replicates itself inside the cell, meaning you need a much smaller dose for the same effect. Lower cost, fewer side effects, easier to scale.
The mRNA that makes more mRNA. It's the gift that keeps on giving, in a carefully controlled way.
ARCT-154, a self-amplifying RNA COVID vaccine from Arcturus Therapeutics and CSL, was approved in Japan in late twenty twenty-three. It's already in use.
I didn't know that had actually reached approval. So the platform is already in deployment, not just in papers.
Then there's the nanoparticle frontier. Ferritin-based nanoparticles that display antigens in a highly ordered array, mimicking the repetitive surface structure of a virus. The immune system has evolved to respond strongly to these repetitive patterns. The NIH's Vaccine Research Center has a ferritin nanoparticle flu vaccine in trials. Early data looks very promising for breadth of protection.
It's almost like we're learning to speak the immune system's language more fluently. Instead of shouting "pathogen!" with a killed virus, we're whispering very specific instructions.
The fifth frontier is what I'd call computational vaccinology. Using machine learning to predict which antigens will be most immunogenic, designing antigens de novo that don't exist in nature but present the right shape to the immune system. This is being applied to HIV vaccine development, which has been the hardest problem in vaccinology for forty years.
Now there's the benchmark. Why has HIV been so hard?
The virus integrates into the host genome and establishes latency within days of infection. The envelope protein is heavily glycosylated — it's coated in sugars that shield it from antibodies. And it mutates at an extraordinary rate, even within a single infected person. The diversity of HIV strains globally dwarfs what we see with influenza or SARS-CoV-2. A vaccine that works against one strain might be useless against another.
You've got a target that hides, mutates, and wears a sugar coat. It's the immune evasion trifecta.
Yet there's been progress. The discovery of broadly neutralizing antibodies — bNAbs — in some HIV-infected individuals showed that the human immune system can produce antibodies that neutralize diverse HIV strains. The problem is those antibodies are rare, take years to develop, and require a specific sequence of B-cell maturation. The vaccine challenge is to guide the immune system through that sequence step by step. It's called "germline targeting." You prime the immune system with an engineered antigen that activates the right B-cell precursors, then boost with a series of slightly different antigens that shepherd the antibody response toward breadth.
That's an entire curriculum. The immune system has to go to school.
Some of the most exciting work is in that space. The Scripps Research Institute and IAVI have a candidate that completed phase one trials with promising results. It's the first time a vaccine has been shown to induce the right B-cell precursors in humans. Still years away from a deployable vaccine, but the principle has been validated.
We've got universal vaccines, therapeutic cancer vaccines, needle-free delivery, self-amplifying RNA, nanoparticle display, computational antigen design, and germline-targeting for HIV. That's seven frontiers. And I'm probably missing some.
You're missing the adjuvant revolution. Adjuvants are the ingredients added to vaccines to boost the immune response. For decades, aluminum salts — alum — were basically the only adjuvant approved in the US. They work fine for antibody responses but don't drive the T-cell responses you need for intracellular pathogens. Now we have a whole toolkit: oil-in-water emulsions like MF59, TLR agonists that directly stimulate innate immune receptors, saponin-based adjuvants like those in the Shingrix shingles vaccine. Shingrix is more than ninety percent effective at preventing shingles in older adults, which is remarkable for a vaccine in an aging immune system. The adjuvant is a huge part of that.
The Shingrix number is stunning. Over ninety percent in older adults. That's a population where vaccines typically underperform because the immune system is less responsive.
It's the adjuvant that makes the difference. The AS01 adjuvant system in Shingrix combines a saponin molecule and a TLR4 agonist. It basically lights up the innate immune system and says "pay attention, this matters.
The adjuvant is the hype man for the antigen.
That's a perfect Corn-ism. The hype man for the antigen. I'm going to use that in my DJ sets somehow.
Please don't. So let me pull back and ask the structural question. You've described five platform eras and at least eight active frontiers. What's driving the acceleration? Is it just more funding post-COVID, or is something deeper happening?
It's multiple things converging. COVID demonstrated that mRNA technology works at scale and can be deployed rapidly — that de-risked the whole field for investors and regulators. But the deeper drivers are scientific. Cryo-EM gives us atomic-resolution structures. Next-generation sequencing gives us pathogen genomes and human immune repertoires. Machine learning lets us predict protein folding and antigen design. Synthetic biology lets us manufacture nucleic acids cheaply and at scale. And our understanding of immunology — the innate immune system, trained immunity, mucosal immunity, the germinal center reaction — has deepened enormously.
It's not just one breakthrough. It's a stack. Better imaging, better sequencing, better computation, better manufacturing, better biology. Each layer amplifies the others.
The regulatory environment has evolved too. The concept of "platform licensure" is being discussed seriously. Instead of licensing each vaccine de novo, you license the platform — say, an mRNA-lipid nanoparticle platform — and then new vaccines using that platform get an accelerated pathway. You just swap in the new sequence. That would dramatically reduce the time and cost of developing vaccines for emerging pathogens.
Which is exactly what you'd want for pandemic preparedness. If a novel virus emerges, you don't want to spend eighteen months on a full clinical development program. You want to sequence the pathogen, synthesize the mRNA, and be in arms within weeks or months.
CEPI — the Coalition for Epidemic Preparedness Innovations — has a hundred-day mission. The goal is to have a vaccine ready for initial authorization within a hundred days of a pandemic pathogen being sequenced. That's down from the three hundred twenty-six days it took for the first COVID vaccine to receive emergency authorization. Which was itself a record.
A hundred days. That's the ambition. What's the realistic timeline for getting there?
It depends on how much of the platform licensure framework gets adopted, how much manufacturing capacity is maintained between pandemics — which is a political and economic question, not a scientific one — and how well prototype vaccines for priority pathogen families work in early trials. CEPI is targeting prototype vaccines for representative viruses from about twenty-five viral families, so that when a new virus emerges, you don't start from zero. You've already done the proof-of-concept work on a related virus and you know what antigen design works.
That's the immunological equivalent of keeping the factory warm. Don't mothball the production line, maintain it in standby.
That's the part that keeps me up at night, honestly. The science is there. The question is whether the political and economic systems will sustain the investment between crises. We have a pattern in public health: panic, spend, forget, get caught unprepared, repeat. COVID cost the global economy something like twelve to fourteen trillion dollars by most estimates. The investment needed to maintain pandemic preparedness is a rounding error on that figure. But it's hard to sustain political attention when there's no visible emergency.
The maintenance budget problem. Nobody wants to pay for roof repairs when it's not raining.
And it's raining less now, metaphorically, so the urgency is fading. But the virological reality hasn't changed. There are hundreds of thousands of mammalian viruses with zoonotic potential. We've barely catalogued a fraction of them.
Let's go back to the number question for a moment. You said twenty-five to thirty diseases have approved vaccines. How many more are in active development?
The WHO's vaccine pipeline tracker lists over a hundred and twenty vaccine candidates in clinical development, targeting diseases that don't yet have licensed vaccines. That includes diseases like respiratory syncytial virus — wait, RSV vaccines were actually approved in twenty twenty-three, so that one just moved from the pipeline to the licensed column. But we're talking about targets like cytomegalovirus, Epstein-Barr virus, norovirus, C. difficile, Group B streptococcus, and of course HIV, tuberculosis, and malaria.
Malaria already has a vaccine though, right?
RTS,S was approved in twenty twenty-one. R21/Matrix-M was approved in twenty twenty-three and is more efficacious and easier to manufacture. Both are about forty percent effective at preventing severe malaria, which is far below the ninety-plus percent we expect from most vaccines, but in a disease that kills over six hundred thousand people a year — mostly children under five in Africa — forty percent is a huge absolute benefit. And the next generation of malaria vaccines using mRNA and whole-sporozoite approaches is already in trials.
Forty percent efficacy would be a failure for a measles vaccine. But for malaria, it's six hundred thousand deaths a year, so forty percent is a quarter of a million lives. Context is everything.
That's the thing about vaccines as a field. The metrics are completely different for different diseases. For a pandemic respiratory virus, you want transmission-blocking immunity and you want it fast. For a slow-progressing chronic infection like HPV, you want durable antibody responses over decades. For a parasitic disease like malaria, you're dealing with a eukaryotic pathogen that has a complex life cycle and sophisticated immune evasion. No single platform solves all of these.
Which is why the frontier is so broad. You need the whole toolkit.
The toolkit keeps expanding. I haven't even mentioned the work on vaccines for non-communicable diseases. Alzheimer's vaccines targeting amyloid-beta and tau are in trials. Vaccines for opioid addiction that generate antibodies against fentanyl, preventing the drug from crossing the blood-brain barrier. Vaccines for hypertension targeting angiotensin. These are all in early-stage trials, but they represent a conceptual expansion of what a vaccine can be.
A vaccine for opioid addiction is a wild concept. You're vaccinating someone against a molecule their own body doesn't produce, so that if they relapse, the drug doesn't reach the brain. It's not preventing infection — it's preventing a behavior from having its pharmacological effect.
It raises enormous ethical questions about consent, about whether you're taking away someone's agency, about who gets to decide. But from a purely technological standpoint, it's remarkable. The immune system can be trained to recognize and neutralize almost any molecule. We're just beginning to explore what that means beyond infectious disease.
The immune system as a general-purpose targeting system. Point it at whatever you want neutralized.
With enormous care, because the immune system is also the thing that gives you autoimmune disease and allergies when it gets it wrong. But yes, the principle holds.
Let's summarize the arc for the listener who's been taking notes. Vaccines started with variolation — controlled exposure to the real pathogen. Jenner introduced the idea of using a related, milder pathogen. Pasteur figured out how to weaken pathogens in the lab. Cell culture unlocked the mid-century explosion of viral vaccines. Recombinant DNA let us make antigens in yeast instead of in pathogens. Conjugate technology solved the pediatric polysaccharide problem. And now we're in the genetic platform era — mRNA, viral vectors — with at least seven active frontiers: universal vaccines, therapeutic cancer vaccines, needle-free delivery, self-amplifying RNA, nanoparticle display, computational design, and germline-targeting for HIV.
Plus the adjuvant revolution and the organizational innovations like CEPI's hundred-day mission. It's not one story, it's a dozen overlapping stories. And the number of vaccines — twenty-five to thirty licensed diseases, over a hundred and twenty candidates in clinical trials targeting new diseases — is a snapshot that will be out of date within a year or two.
The question behind the question was "is vaccine technology still evolving." I think the answer is: it's evolving faster than at any point since Jenner. Maybe faster than at any point, period.
I'd go further. We're witnessing the transition from vaccines as a craft — grow it, kill it, inject it — to vaccines as a design discipline. We're learning to specify immunological outcomes at the molecular level. That's not an incremental improvement on what came before. That's a phase change.
The craft era still produced smallpox eradication. One of the greatest achievements in human history, done with what we'd now consider primitive technology. So the fact that the tools are getting better doesn't diminish what came before. It just raises the ceiling on what's possible.
Smallpox eradication is worth pausing on for a second. Declared in nineteen eighty. A disease that killed an estimated three hundred million people in the twentieth century alone — gone. No cases since nineteen seventy-seven. That was done with a live vaccinia virus vaccine, bifurcated needles, and a global surveillance and containment strategy. No mRNA, no nanoparticles, no cryo-EM. Just extraordinary organization and political will.
Now we have the organization, the political will — sometimes — and the molecular tools. The combination should be unstoppable. The question is whether we use it.
That's the open question. The technology is not the limiting factor anymore. It's us.
And now: Hilbert's daily fun fact.
Hilbert: In nineteen thirty-five, the Icelandic naturalist Bjarni Sæmundsson documented the smallest butterfly ever recorded in Iceland — the Greenland sulphur, whose wing scales are structured at the nanoscale to absorb heat so efficiently that the butterfly can remain active at temperatures just above freezing, a feat no other butterfly in the subarctic can match.
A butterfly with nanotech heating elements. Of course there is.
The question that follows from all this — if we're in a phase change, if the ceiling is higher than ever — is what vaccine should exist in twenty years that doesn't exist today? What's the aspirational target?
A pan-coronavirus vaccine that protects against all future spillovers. A durable HIV vaccine. A truly universal flu vaccine. A therapeutic cancer vaccine that's as routine as a tetanus shot. A malaria vaccine that hits ninety percent efficacy. And — this is the big one — a platform that can go from pathogen sequence to authorized vaccine in under a hundred days, every time, without panic.
That's a list worth working toward. And it's not science fiction. Every item on that list has a prototype in development or a clear technical path.
The path is there. The question is whether we walk it.
This has been My Weird Prompts. Thanks to our producer, Hilbert Flumingtop, for the fact-checking and the subarctic butterfly trivia. If you enjoyed this episode, leave us a review wherever you get your podcasts — it helps other people find the show. We'll be back next week.
See you then.
