How to fix science

here’s a widespread understanding that science is in a slump. To many people, the solution is obvious: clear out the ‘deadwood’ and create a program to find and reward young geniuses. Indeed it’s true that early-stage investigators have trouble getting funded. But the idea of ‘genius grants’ is based on a myth. The idea is that scientists are most productive while young, so we should single out the brightest ones and let them apply for a special grant limited to, say, people under thirty. To nonscientists that might sound like a great idea: find the geniuses and give them what they need. But it won’t work.

Oh God, not another special grant program

First, the idea that creativity declines with age is not true, not just in biomedicine but in general. If they seem less creative, it is because they’re forced to spend most of their time writing grants and useless papers. Only one in ten grants get funded and it takes three to four months of work at seventy hours a week to write one. You can only revise and resubmit it once. The project must be exciting to the reviewers, which means it must conform to what they like and what techniques they think you must do. You don’t need to be a genius to know this kills imagination. You never ever put an imaginative idea in a grant—it would kill it dead.

Graveyard of a failed experiment

A new grant program would just push young scientists onto the same funding treadmill that is killing scientific productivity among us geezers. I’ve seen it in dozens of grants. Enormous creativity is evident in the amount of trickery they use. Most applicants are lying: either they plan to say the idea didn’t work so they can study the gene they really want; or they have already done the work they propose, which is how they can be so sure it will work. This is not cynicism. It is advice that was given to me: write your grant to propose what you have already done and use the money to do the next thing.

Even if NIH handed out grants to early stage investigators like candy, universities would simply adapt by hiring young researchers and then replacing them when they encounter an obstacle. That happens often to early investigators who don’t know their idea has already been tried by someone else and abandoned. They also haven’t discovered the myriad ways Nature has of fooling them. To them every little smudge on a Western blot is a big discovery and every new idea is a cure.

Early researchers need the flexibility to try new ideas and change them. The solution is not to force them to write grants all the time, which is what a special grant program would do, but to continue independent research well beyond grad school in a real environment. For that they need money. But they also need the flexibility to drop an idea when it doesn’t work. Grants would make this harder. The solution is to just give them a start-up fund, as universities used to do. Just make sure the school doesn’t get its paws on it, because they’ll take it for themselves.

The myth of age-related loss of creativity

A famous physicist said once said the old ideas won’t die until the old ones who believe them die off. He meant that people become invested in making the old theory work. They see the flaws in the new one and their career would suffer if the old one turned out to be wrong, so they resist the new theory.

But the idea somehow turned into a slur, convincing people that scientists lose their creative ability with age. Modern neuroscience has shown again and again that this doesn’t happen until disease or advanced age take a toll on the brain. It does not mean that people in their twenties are inherently more or less capable of inventing new ideas. What it means is that the system weeds out anyone who does. Anyone with a new, unconventional idea is forced out.

As airline pilots say, there are old pilots and bold pilots, but no old bold pilots. So it is in science. It is a form of natural selection: those most fitted for the stagnant bureaucracy that is academia are the ones who stay there.

People often mention Einstein, who invented the theory of relativity in his twenties, as a counter­example. But Einstein was not employed in academia or in any formal lab at the time. After he was granted an academic position for his success, his productivity (we are told) plummeted. This was not due to age. It was a searing indictment of academia of his time, and of our time as well.

Some of the most skilled and inventive young researchers I’ve known were forced out of science. One became vice president of Pfizer International the next day. Another one, tossed out for budget­ary reasons, became head of a big independent research institute two weeks later. We felt sorry for them: the poor devils died and became bureaucrats.

Most early ideas in biomedicine suck on ice

The second problem, in my observation of students and postdocs, is that their ideas are generally not as brilliant as everybody thinks. Good ideas don’t just happen: they evolve in an ecosystem full of misinformation, misleading clues, and missing pieces. Genius grants might work in mathematics, where everything is logical, true, and discoverable. But in biomedicine it would fail.

The NIH is already well aware that early-stage investigators (NIH specifically uses the term ‘early’, not ‘young’, for good reasons) suffer from many disadvantages: they lack preliminary results to generate enthusiasm among grant reviewers. They don’t have the grantsmanship skill (i.e. deviousness) needed to make the screwed-up system work. Worst of all, the face the greedy policy universal in academia today: no active grant, no job interview. Essentially they have to bribe the school to hire them. This is the real problem in science and a few genius grants won’t solve it.

NIH already gives early investigators a huge break in grant scoring. The cutoff score for one’s first R01 is five to ten points higher than for everyone else. Grants are scored on a percentile. A lower score is better. Normally only a score below 20 is considered for funding, so only the top 20% have a chance. Raising it to 25 gives them a tremendous advantage. It’s not a linear curve, so it’s an even bigger advantage than it looks.

The myth of the reproducibility crisis

Another symptom of the decline of science is the so-called reproducibility crisis. I’ve written before that it is misnamed. It is actually a competency crisis. I once tried teaching a group the correct way to analyze their images. They’d been doing it incorrectly for years. Repeatedly they asked me to tweak their images so they could get better results. I had to explain over and over that this would be fraud and would yield irreproducible results. They didn’t care. They were getting grants and that’s all that mattered. The rest of it was unimportant.

I’ve seen other labs with smart people who were doing the same thing. I won’t bore you with any more anecdotes, but when I was dragged into academia it confirmed my worst stereotypes about universities. The vast majority of academics only care about one thing: funding. The greed in their eyes when they discuss it is scary. They will do anything, screw over anybody, lie, and fabricate data if that’s what it takes. Most of them might prefer to do things honestly, but that is not what the system rewards.

Solutions

Here is what must happen if we really want to restore science.

1. Eliminate Program Announcements at the Center for Scientific Review (CSR). They don’t work. They act as barriers, dissuading academics from considering unorthodox ideas. Researchers read them to find out what NIH wants, then write a grant to give it to them. If there’s no specific PA, your only alternative is to change to a less interesting topic or dump the grant into the ‘general’ category where it will never be funded.

2. Give all early stage investigators a start-up fund to allow them to compete fairly for positions. Cutting grant overhead costs, which typically add 50% to each grant for the bureaucrats, would easily pay for it.

3. Remove grants and papers from the criteria for productivity. People should not have to spend all their time writing grants and junk papers instead of making discoveries. The easiest way to do this is to remove science from academia altogether. A new type of institution would attract young researchers and encourage others to pursue important projects that demand long-term effort. It would be flooded with applicants who are tired of being terrorized by power-mad academic bureaucrats. Let the schools teach if they can.

4. Teach some boring but important stuff like data collection, statistical analysis, image forensics, and data archival at the graduate schools. This would virtually eliminate the ‘reproducibility crisis.’

There are other things that would help: Find a way for negative results and failed ideas to be published. This will prevent young researchers from trying things that are known not to work. Journals dedicated to this topic have mostly failed. And continue making more scientific journals, including old ones, free to the public that paid for the work. This will help avoid replication and build scientific literacy among the public, which translates into stronger support.

Jay Bhattacharya may be America’s last chance for revitalizing biomedical research. It’s now too late to hope to maintain parity with China, and some of the changes will not be easy. The corrupt system is the source of the bureaucrats’ power to terrorize faculty; the academics are too risk-averse; and the science journals are still too politicized. Yet change is essential. You heard it from this old geezer first.

jun 11 2026, 5:57 am