Daniel sent us this one — he's asking why we've had antibiotics for decades but still can't treat routine viral illnesses like the common cold. We can cure strep throat in a week, but for a cold we're still saying rest and hydrate. What makes viruses so much harder to target? And is it just biology, or are there economic and diagnostic barriers too?
I'm Herman Poppleberry, this is my brother Corn, and this is a question that I think frustrates people more than they realize. You go to the doctor with a cold, you feel terrible, and the best medical science can offer is orange juice and a nap. It feels like failure.
It feels like the doctor is basically saying, good luck with that.
But the reasons are genuinely fascinating, and they go way deeper than "we just haven't tried hard enough." Let's start with the fundamental difference between a bacterium and a virus, because that's where the whole asymmetry begins.
Why exactly can we wipe out a bacterial infection in a week, but a cold just has to run its course?
It comes down to what these two things actually are. Bacteria are free-living organisms. They're what biologists call prokaryotes — they have their own cell machinery, their own metabolism, their own ribosomes for making proteins. A bacterium is a complete, self-contained factory. A virus is not. A virus is essentially genetic material — DNA or RNA — wrapped in a protein coat, sometimes with a lipid envelope. It has no metabolism, no ribosomes, no ability to make energy or proteins on its own. It's a piece of genetic information that can only reproduce by hijacking your cells.
Bacteria are like independent contractors who show up with their own tools and get to work. Viruses are squatters who move into your house and start using your furniture.
That's exactly the right analogy. And that distinction is everything. When you want to kill a bacterium, you have a whole menu of targets that exist in the bacterium but have no equivalent in human cells. The classic example is the peptidoglycan cell wall. Bacteria have this rigid mesh-like structure outside their cell membrane that maintains their shape and prevents them from bursting. Human cells don't have anything like it. Penicillin and its relatives — the beta-lactam antibiotics — bind to proteins the bacteria use to cross-link that peptidoglycan mesh. The wall weakens, the bacterium pops, and you're fine because your cells never had that target in the first place.
It's like finding a lock on the contractor's toolbox that doesn't exist anywhere in your house. You break the lock, the contractor's tools fall apart, your furniture is untouched.
And it's not just the cell wall. Bacterial ribosomes — the molecular machines that translate genetic code into proteins — are what's called 70S ribosomes, made of a 50S and a 30S subunit. Human ribosomes are 80S, with 60S and 40S subunits. They're different enough that drugs like tetracyclines, macrolides, and aminoglycosides can bind to the bacterial version and block protein synthesis while leaving your ribosomes alone.
We've got multiple targets that are completely foreign to human biology. The bacterium is full of them.
Bacterial DNA gyrase — no human equivalent, target for fluoroquinolones. The bacterial metabolic pathways for synthesizing folic acid — humans get folic acid from diet, bacteria have to make it, so sulfonamides and trimethoprim block those enzymes. Each of these is a clean, specific target. Now let's flip to viruses. A virus enters your cell, sheds its coat, and the viral genetic material essentially says to your cellular machinery: stop what you're doing, make copies of me instead.
At that point, the virus is using human enzymes to replicate. If you try to poison the replication process, you're poisoning your own cells.
That's the core problem. Most steps in the viral life cycle are performed by host enzymes. The virus might bring along one or two of its own — a polymerase for copying its genome, a protease for processing its proteins — but compared to bacteria, the number of virus-specific targets is vanishingly small. And the targets that do exist are often not conserved across different viruses. The enzyme that replicates influenza's genome is completely different from the one that replicates rhinovirus.
Even if you find a target, you're building a drug for one virus, not a broad-spectrum antiviral.
And here's the thing — we do have some antivirals, and they work by targeting those rare viral-specific enzymes. Let me walk through a few so we can see the pattern. HIV's reverse transcriptase is a beautiful example. Normal human biology never converts RNA into DNA — that's not a thing our cells do. But HIV is a retrovirus, and it carries an enzyme that does exactly that: it reads the viral RNA and writes it into DNA that can integrate into your genome. That enzyme, reverse transcriptase, has no human counterpart. Drugs like zidovudine and tenofovir target it, and they work.
A clean lock to break.
A clean lock. Same with HIV protease — a viral enzyme that cuts long protein chains into functional pieces. Drugs like ritonavir and darunavir bind to that protease. The host cell doesn't use that protease for anything. Influenza has a different set of targets. The virus has two proteins on its surface: hemagglutinin, which helps it enter cells, and neuraminidase, which helps newly made viruses cut themselves free from the infected cell so they can go infect other cells. Oseltamivir — Tamiflu — and zanamivir block neuraminidase.
Which is why Tamiflu actually works, but only if you take it early. Once the virus has already spread through your respiratory tract, blocking the exit mechanism for new viruses is closing the barn door after the horses are gone.
That timing problem is huge with antivirals, and we'll come back to it. But let me give you one more example — remdesivir. It's a nucleotide analog that targets viral RNA-dependent RNA polymerases. These are enzymes that RNA viruses use to copy their RNA genomes. Human cells don't have RNA-dependent RNA polymerases, so again, it's a viral-specific target. Remdesivir was originally developed for Ebola, failed in those trials, and then got repurposed for COVID-19.
Which is a theme, apparently. Antivirals that fail for one virus get dusted off for the next pandemic.
Because the pipeline is so thin. And that brings us to the next layer of difficulty. Even when we find a viral target, the virus fights back fast. RNA viruses — influenza, rhinovirus, hepatitis C, coronaviruses — have mutation rates around ten to the minus four to ten to the minus five per nucleotide per replication cycle. Bacteria mutate at about ten to the minus eight. That's a difference of roughly ten thousand to a million times.
The virus is not just a moving target — it's a target that's actively shape-shifting while you're trying to aim.
It's not just the rate. It's the population dynamics. A viral infection isn't one genome replicating. It's what virologists call a quasispecies — a cloud of slightly different genetic variants all coexisting in the same host. Most of them are less fit than the dominant strain, but some of them, purely by chance, carry mutations that confer resistance to whatever drug you're using. When you apply the drug, you kill off the dominant, drug-sensitive strain. The resistant mutants, which were maybe one in a million, suddenly have no competition.
It's like trying to weed a garden where the weeds evolve herbicide resistance before you finish spraying.
This is exactly what happened with the first generation of influenza antivirals. Adamantanes — amantadine and rimantadine — were introduced in the 1960s and targeted the M2 ion channel of influenza A. By the 2005-2006 flu season, over ninety percent of circulating H3N2 strains were resistant. They're essentially useless now.
The target pool is tiny, and the targets that do exist are constantly mutating. But there's another piece here that I think gets overlooked — the diagnostic problem. When you have strep throat, the doctor can do a rapid strep test, confirm it's bacterial, and prescribe penicillin. When you have a cold, nobody tests to find out which of the two hundred rhinovirus serotypes you have.
More than two hundred serotypes of rhinovirus alone. And that's just rhinovirus. The common cold is also caused by coronaviruses, adenoviruses, respiratory syncytial virus, parainfluenza virus, human metapneumovirus. It's a syndromic diagnosis — we call it a cold because of the symptoms, not because we know what's causing it.
Even if we had a drug for rhinovirus type fourteen, nobody's going to run a PCR panel, wait two days for results, and then prescribe it for a condition that's already half over.
Bacterial infections can be cultured in a lab — you grow them on agar plates and identify the species in twenty-four to forty-eight hours, sometimes faster with rapid tests. Viral diagnosis requires PCR or antigen testing, and for mild illnesses, it's simply not done. Without a specific diagnosis, you can't use a targeted antiviral. And without the ability to use targeted antivirals, you're stuck with broad-spectrum guesses, which is a recipe for resistance and side effects.
We've got three layers so far. Layer one: viruses have almost no unique targets compared to bacteria. Layer two: the targets that do exist mutate at blinding speed. Layer three: we can't even diagnose which virus someone has in time for treatment to matter. Is there a layer four?
Layer four is money. The economics of antiviral development are brutal in ways that most people don't appreciate. The estimated cost to bring a new drug from discovery to market is about two point six billion dollars. For a bacterial infection, you're treating something that might land someone in the hospital, cause sepsis, require weeks of IV antibiotics. There's a willingness to pay. For a cold, you're treating a five-day illness that people manage with over-the-counter symptom relief. Nobody is going to pay five hundred dollars for a cold pill.
Even if they would, by the time they realize they're sick, get a prescription, and fill it, the cold is already peaking. The window for an antiviral to make a meaningful difference is maybe the first twenty-four to forty-eight hours.
That's exactly what we saw with Paxlovid for COVID-19. Nirmatrelvir, the active component, is a protease inhibitor that blocks the SARS-CoV-2 main protease — again, a viral-specific enzyme. In clinical trials, when given within five days of symptom onset, it reduced hospitalization or death by about eighty-nine percent in high-risk patients. That's a effective antiviral.
The real-world impact has been limited because you need a positive test, a prescription, and access to the drug within that five-day window. A lot of people don't test until day three or four, or they test negative on a rapid test early on, or they can't get a doctor's appointment in time.
The logistical hurdles eat into the clinical benefit. And that's for a virus that caused a global pandemic with enormous resources thrown at it. Now imagine trying to build that infrastructure for the common cold. It's not going to happen.
The market for acute antivirals is fundamentally broken. The patients who need them won't pay enough, and the window for treatment is too narrow to build a business case around.
Here's the exception that proves the rule: chronic viral infections are a completely different story. HIV and hepatitis C antivirals are among the most successful drugs in history. Because the infection doesn't go away on its own. Patients will take a daily pill for life, or in the case of hepatitis C, a twelve-week course that costs tens of thousands of dollars. The market is enormous, the treatment window is indefinite, and the clinical benefit is unambiguous.
The economics work for chronic infections, which represent a tiny fraction of viral illnesses, and they fail for acute infections, which represent the vast majority of what people actually get.
The numbers bear this out. The global antibiotics market was around forty billion dollars in 2024. The antivirals market, excluding HIV drugs, was around fifteen billion. Viral infections are far more common than bacterial infections, but the market is less than half the size.
There's a specific case study I want to dig into, because it illustrates all of these problems in one story.
Oh, this is a great and tragic example. Pleconaril was developed in the late 1990s and early 2000s as an anti-rhinovirus drug. It worked by binding to a pocket in the rhinovirus capsid — the protein shell — and preventing the virus from uncoating and releasing its RNA into the host cell. It was effective in clinical trials. It reduced cold symptoms by about a day, which doesn't sound like much but is actually meaningful for a self-limiting illness.
The FDA rejected it.
The FDA rejected it in 2002. First, it caused liver enzyme elevations in some patients — a signal of potential liver toxicity. Second, and this is the detail that really captures the practical difficulty, it interacted with hormonal birth control. Pleconaril induced the CYP3A4 enzyme in the liver, which metabolizes estrogen, making oral contraceptives less effective. For a drug you'd be prescribing to women of reproductive age for a non-life-threatening condition, that's a deal-breaker.
You had a drug that worked, for a condition everyone gets, and it failed because of side effects that would be totally acceptable for a cancer drug but are unacceptable for a cold.
The risk-benefit calculus is completely different. For a life-threatening bacterial infection, we accept significant toxicity. Gentamicin can cause kidney damage and hearing loss, but we use it because the alternative is death from sepsis. For a cold, the bar for safety is essentially zero — the condition resolves on its own, so any serious side effect is unacceptable.
Which means antiviral development for mild illnesses is playing the game on the hardest difficulty setting. You need a drug that's extremely safe, effective across multiple serotypes, and cheap enough that people will pay for it out of pocket, for a condition that lasts less than a week.
You need to find a target that is conserved across all those serotypes, essential for viral replication, and absent from human biology. That's a needle in a haystack, and the haystack is on fire and moving.
Let me ask you something about the hepatitis C story, because you mentioned it as an exception. What made sofosbuvir work when so many other attempts failed?
Sofosbuvir and the other direct-acting antivirals for hepatitis C are one of the great success stories in modern medicine. Hepatitis C has an RNA-dependent RNA polymerase called NS5B, which is relatively stable across genotypes. Not perfectly stable — there are differences between genotypes one through six — but stable enough that a nucleotide analog like sofosbuvir can bind to the active site and cause chain termination, stopping the virus from copying its genome. The key differences from rhinovirus are, first, the polymerase is more conserved. Second, the infection is chronic, so you have months to treat, not days. Third, the consequences of not treating — cirrhosis, liver cancer, liver failure — are catastrophic, so the risk-benefit calculus favors aggressive treatment.
Fourth, the price tag was eighty-four thousand dollars for a twelve-week course when it launched, which pharma companies noticed.
The price has come down significantly with generics and competition, but yes. Sofosbuvir made Gilead over ten billion dollars in its first year. That's the kind of return that justifies the two-point-six-billion-dollar development cost. No cold drug will ever generate that kind of revenue.
We've established that the antiviral gap is biological, evolutionary, diagnostic, and economic. But I want to push on one thing. You mentioned earlier that bacteria mutate slower. They still mutate, and antibiotic resistance is a global crisis. If bacteria can evolve resistance to dozens of antibiotics over decades, why haven't we lost antibiotics the way we've lost some antivirals?
This is where the target count really matters. Bacteria have evolved resistance to virtually every antibiotic we've thrown at them, but we had dozens of targets to begin with. When penicillin resistance emerged through beta-lactamases — enzymes that break the beta-lactam ring — we developed beta-lactamase inhibitors like clavulanic acid. When that wasn't enough, we moved to cephalosporins, carbapenems, completely different beta-lactam structures. When methicillin-resistant staph aureus emerged, we had vancomycin, linezolid, daptomycin. Each of these hits a different target.
The bacterial target landscape is like a large apartment building with many doors. If the bacteria learn to lock one door, you try another. The viral target landscape is a studio apartment with maybe two doors, and one of them is on fire.
The other one keeps changing its lock. That's not even an exaggeration. For influenza, we basically have two target classes — neuraminidase inhibitors and the older adamantanes, which are already useless. For rhinovirus, we have nothing approved. For adenovirus, nothing. For norovirus — the stomach flu that tears through cruise ships — nothing approved.
The stomach flu one is particularly interesting because norovirus is incredibly contagious, causes enormous economic disruption, and we have zero pharmaceutical tools for it.
Norovirus doesn't even grow well in cell culture, which makes drug discovery extremely difficult. You can't screen compounds if you can't grow the virus in a dish. For decades, the only way to study norovirus was to infect human volunteers, which is ethically complicated and expensive. They finally developed a cell culture system — using human intestinal organoids, essentially mini-guts in a dish — but that was only in 2016.
We're talking about a virus that infects nearly everyone multiple times in their life, causes about two hundred thousand deaths globally per year, mostly in children and the elderly in developing countries, and we couldn't even grow it in a lab until eight years ago.
Ten years ago, right. And that's another dimension of the problem. Bacteria are comparatively easy to study. You can grow most pathogenic bacteria on agar plates, test antibiotics against them, and get results in days. Viruses require host cells to replicate, so you need tissue culture systems, which are finicky and don't exist for every virus.
I want to circle back to something you said about the common cold being caused by multiple viruses. I think there's a widely held misconception that the cold is one thing — one virus, one disease — and if we just put enough money into it, we'd have a cure.
It's not even close to one thing. Rhinoviruses account for maybe thirty to fifty percent of colds. There are over two hundred serotypes of rhinovirus, and they're not closely related enough that one drug or one vaccine can cover all of them. Then you have four endemic human coronaviruses — not SARS or MERS, but the ones that cause fifteen percent of colds. Adenoviruses, RSV, parainfluenza, human metapneumovirus — each of these is a completely different virus with different replication strategies, different enzymes, different structural proteins.
A broad-spectrum antiviral for the common cold would need to hit targets across five or six completely different viral families. That's like asking for a single pesticide that kills beetles, aphids, caterpillars, and fungal blight without harming the plant.
That's why the few broad-spectrum antivirals we do have are things like interferon, which works by stimulating the host immune response rather than targeting the virus directly. But interferon has brutal side effects — flu-like symptoms, fatigue, depression — because you're essentially cranking the immune system into overdrive. It's used for hepatitis C in some cases, but you wouldn't take it for a cold.
Where does this leave us? Let me try to synthesize. The antiviral gap has three pillars. Pillar one is biological: bacteria are free-living organisms with dozens of unique, druggable targets; viruses are genetic parasites with maybe one or two virus-specific enzymes, and those enzymes mutate fast. Pillar two is diagnostic: we can identify bacterial infections quickly and cheaply; we can't do the same for most viruses, and without a diagnosis, targeted treatment is impossible. Pillar three is economic: the market for acute antivirals is broken because the treatment window is narrow, the illness is self-limiting, and the safety bar is impossibly high.
I'd add a fourth pillar, which is historical. The golden age of antibiotic discovery — roughly the 1940s through the 1960s — was fueled by screening soil microbes. Streptomycin, tetracycline, vancomycin, erythromycin — all came from soil bacteria, mostly Streptomyces species. You could literally dig up dirt from different parts of the world, culture the microbes, and test them against pathogens. It was a low-cost, high-yield strategy that produced dozens of drug classes.
The equivalent for antivirals doesn't really exist. You can't screen soil microbes for antiviral activity in the same way because viruses need host cells to grow.
The antiviral discovery process has been much more rational and targeted — identify a viral enzyme, solve its crystal structure, design a molecule that fits the active site. That's slower, more expensive, and has a higher failure rate. The first antiviral, idoxuridine, was approved in 1963 for herpes keratitis — that's thirty-five years after penicillin was discovered in 1928. And it was a topical treatment, not a systemic drug.
Thirty-five years to get from the first antibiotic to the first antiviral, and the antiviral was essentially eye drops. That really frames the difficulty.
Even now, almost a century after Fleming's discovery, we have over a hundred approved antibiotics across more than a dozen classes, and maybe thirty to forty approved antivirals total, mostly for HIV, hepatitis, and herpesviruses.
When someone says "why can't we cure the common cold," the honest answer is: we've been trying for sixty years, the biology is harder, the economics don't work, and the diagnostic infrastructure doesn't exist. It's not that medical science is lazy or indifferent. It's that the problem is a perfect storm of constraints.
I think that's worth sitting with, because it reframes the frustration. When your doctor says rest and hydrate, that's not medical nihilism. It's a rational acknowledgment that the tools don't exist, and the ones that might exist someday aren't here yet.
The practical takeaway is that symptom management is currently the best we've got for most viral illnesses, and it's not because nobody cares. It's because viruses have evolved to be exceptionally difficult to drug, and our economic system isn't set up to solve problems where the treatment window is five days and the patient won't pay more than twenty dollars.
There's a policy dimension here too. The market failure for antibiotics is well documented — we've talked about that before, the problem where companies don't invest because new antibiotics are used sparingly to preserve effectiveness. The market failure for antivirals is different but related. For acute viral infections, the commercial incentive barely exists. That's where public funding comes in. BARDA — the Biomedical Advanced Research and Development Authority in the US — funded remdesivir's development for Ebola and later COVID-19. Without that, the drug probably wouldn't exist.
Pandemic preparedness isn't just about stockpiling masks and vaccines. It's about funding antiviral research for viral families that have pandemic potential — coronaviruses, influenza, paramyxoviruses — even when there's no immediate market.
The frustrating thing is, we know which viral families pose the greatest risk. The WHO has a blueprint list of priority pathogens. But funding is episodic — it surges during a pandemic and evaporates afterward. We saw this with SARS in 2003, with Ebola in 2014, with Zika in 2016. Each time, research funding spiked and then collapsed when the immediate threat passed.
The cycle of panic and neglect.
And that's a terrible way to develop drugs that take a decade or more from discovery to approval.
Let me ask you a forward-looking question. You mentioned mRNA technology and CRISPR earlier. Is there a path where those platforms change the antiviral equation?
There's genuine reason for cautious optimism. mRNA vaccines showed us that we can design and manufacture vaccines against novel viruses in under a year. That's transformative. On the antiviral side, there's work on CRISPR-based systems — one approach called PAC-MAN, which stands for Prophylactic Antiviral CRISPR in Human Cells, uses Cas13 to target and degrade viral RNA. The idea is that you could design guide RNAs against conserved regions of a viral genome and deliver the system as a prophylactic or early treatment.
Instead of designing a small molecule to fit a viral enzyme's active site, you're designing an RNA guide to match a viral sequence. The design problem shifts from chemistry to informatics.
In principle, yes. In practice, there are enormous delivery challenges — getting the CRISPR components into the right cells in the respiratory tract, avoiding off-target effects, managing immune responses to the delivery vector. It's promising but not imminent. There's also interest in host-directed therapies — drugs that target host factors the virus needs rather than the virus itself. The advantage is that host targets don't mutate, so resistance is less of a problem. The disadvantage is that you're interfering with normal cellular functions, which brings us back to the toxicity problem.
The squatter's furniture problem again. If you target the host machinery, you're damaging your own house.
But the hope is that you can find host factors that the virus absolutely requires but that the host cell can tolerate losing temporarily. There's some interesting work on this for influenza and dengue.
The future might involve prophylactic CRISPR treatments, host-directed therapies, and mRNA vaccines that can be updated faster than the virus can drift. But we're not there yet, and the fundamental biology hasn't changed — viruses are still squatters with very few tools of their own.
The economics haven't changed either. Unless we create market incentives — advance purchase commitments, prize funds, patent extensions on other drugs in exchange for antiviral development — the pipeline will remain thin.
Next time someone tells you to drink orange juice for a cold, you can explain that it's not medical laziness. It's the rational endpoint of a problem that has stumped biology, chemistry, and economics for sixty years.
Maybe, if we're lucky, our kid or our grandkid will look back at "rest and hydrate" the way we look at leeches and bloodletting. But we're not there yet.
Now: Hilbert's daily fun fact.
Hilbert: The oldest known manuscript to describe fungal mycelial networks is a ninth-century Irish monastic text that compares the underground threads to the interconnectedness of the Holy Trinity, but the scribe also included a marginal note complaining that the same fungi had ruined his breakfast porridge.
The Irish were doing mycology and breakfast reviews in the same manuscript margin.
Some things never change.
This has been My Weird Prompts. Our producer is Hilbert Flumingtop. If you enjoyed this episode, tell someone the real reason we can't cure the common cold — they'll either be fascinated or ask you to pass the tissues. Find us at myweirdprompts.
Until next time.