Tech has a story problem.
I saw a tweet from Sebastian Caliri recently that crystallized something I've been mulling over. The gist: most Americans see AI as job-killer-for-billionaires. The only thing it seems to offer ordinary people is unemployment and more wealth for the already wealthy. We, in the tech community, need a better story.
The data backs this up. According to Pew Research, 51% of Americans are now more concerned than excited about AI - compared to just 15% of AI experts. That's not a gap. That's a chasm. And YouGov polling shows negative sentiment jumped from 34% to 47% in just six months.
Here's the uncomfortable truth: the skeptics aren't entirely wrong.
Yes, much of AI has been sold as "productivity gains" that mostly flow to shareholders. Yes, healthcare costs are bankrupting families - one-third of Americans find their healthcare unaffordable. Yes, job displacement fears are real, even if the numbers so far suggest AI created more jobs than it destroyed in 2024 (roughly 119,900 created versus 12,700 lost, according to ITIF).
People don't care about "competing with China" when they can't afford a house and medical bills are piling up. That critique deserves a real answer, not dismissal.
But here's what's actually happening in the labs - and it's not chatbots.
AI isn't just automating spreadsheets. It's learning to read biology the way AlphaFold reads proteins. And the implications are staggering.
Drug discovery that used to take a decade now takes months. AlphaFold, the AI that solved protein structure prediction - earning its creators a Nobel Prize - has been used to identify drug candidates in 30 days instead of years. Insilico Medicine went from target identification to Phase IIa clinical trials in under three years - a process that traditionally takes 10-15 years.
AI is finding anti-aging compounds with unprecedented accuracy. A May 2025 study published in Aging Cell by researchers at Scripps Research and Gero tested 22 AI-identified compounds in C. elegans (a standard model organism for aging research). Sixteen extended lifespan - a 73% hit rate that's roughly 1,000 times higher than traditional unbiased drug screens. One compound extended lifespan by 74%, among the most effective ever recorded.
Is this worms, not humans? Yes. But the methodology breakthrough matters: AI can now identify compounds that act across multiple aging pathways simultaneously - something human researchers rarely attempt because the complexity is overwhelming.
And it's not just biology. AI is accelerating breakthroughs across every domain where complexity has been the bottleneck - including the energy systems that power everything else.
Fusion energy is no longer "always 30 years away." AI is solving plasma instability problems that stymied physicists for decades. Princeton researchers used reinforcement learning to predict and prevent plasma tearing in real-time. Private fusion funding jumped from $1.7 billion in 2020 to $15 billion in 2025. Commonwealth Fusion Systems has signed power deals with Google for a commercial plant slated for the early 2030s.
Near-free energy changes everything. It changes the cost of computation, desalination, manufacturing, and yes - drug development.
Let me bring this back to something concrete: your medical bills.
Chronic disease costs the US economy $3.7 trillion annually when you include lost productivity - that's roughly one-fifth of GDP. The CDC reports that 86% of healthcare spending goes to patients with chronic conditions. Diabetes alone costs $413 billion per year.
What if we could actually prevent these diseases instead of just managing them?
A 2023 NBER study found that wider AI adoption could reduce US healthcare spending by $200-360 billion annually - without reducing quality of care. A systematic review in Nature Digital Medicine spanning 19 clinical studies confirmed that AI improves diagnostic accuracy while reducing costs, largely by minimizing unnecessary procedures.
This isn't theoretical. Over 500 AI-related drug submissions have reached the FDA since 2016. More than 200 AI-enabled drugs are expected to receive regulatory approval in the next five years.
Here's where I should probably disclose my bias: I've spent the last few years going down the longevity rabbit hole. I've attended the Longevity Biotech Fellowship and Vitalist Bay. I've contributed to Biolearn, an open-source aging clock consortium. I'm building Beginnity, which starts as a fitness app but has longer-term aspirations around biological optimization.
So take my enthusiasm with appropriate skepticism. But here's what I've seen in those rooms:
The researchers working on aging aren't fringe. They're at Scripps, Stanford, Harvard, and well-funded startups like NewLimit (co-founded by Coinbase's Brian Armstrong), Junevity ($20 million seed for cell reprogramming with siRNA, partnered with Novo Nordisk), and BioAge ($170 million Series D in 2024). They're using AI not to incrementally improve existing drugs, but to fundamentally rethink how we target the cellular damage that accumulates with age.
Insilico Medicine's Rentosertib is reportedly the first AI-designed drug with anti-aging applications to reach human trials. It targets TNIK, a protein the AI identified as linked to aging - one that wasn't on most researchers' radar.
The University of Copenhagen is using AI to screen compounds that clear senescent cells from the brain. India's IIT-Delhi developed AgeXtend, an AI platform to identify molecules promoting healthy aging.
These aren't moonshots for 2050. They're Phase II trials now.
I'm not going to pretend this is all sunshine. The first CRISPR therapy, Casgevy, costs $2.2 million per patient. Some gene therapies run $3.5 million. That's not democratized healthcare - that's another luxury for the wealthy.
But unlike SaaS productivity tools - where gains flow disproportionately to employers - curing disease actually helps everyone. The economics of healthcare are different. When we cure sickle cell disease, Medicaid stops paying for lifetime management. When we prevent Alzheimer's, Medicare saves the $360 billion annually projected for dementia care by 2024.
That's why CMS launched the Cell & Gene Therapy Access Model - outcomes-based payment for expensive therapies. And in late 2024, the FDA convened researchers to "platformize CRISPR," creating standardized pathways that could dramatically cut costs for future gene therapies.
The path to democratization isn't automatic. But it exists.
Here's what I think gets missed in most AI discourse: the benefits don't have to stay locked in research labs or Fortune 500 boardrooms.
Robotics is becoming accessible to small businesses. The collaborative robot market is projected to grow from $2 billion to $12 billion by 2030, but the real story is the business model shift. Companies can now rent robots, test automation without massive capital commitments, and scale only when results justify it. AI-enabled robots deliver 40% higher efficiency than traditional automation - and they're increasingly targeting the repetitive, physically demanding work that burns people out.
Agriculture is getting the same treatment. The agricultural robotics market is growing at 24.7% annually - more than double the overall robotics market. What matters is how: subscription-based services and modular platforms are making precision farming accessible to smallholders, not just industrial operations. That's how you actually reduce food costs.
Healthcare AI is reaching underserved areas. Diagnostic tools that were once exclusive to advanced hospitals are now deployable in remote clinics. AI doesn't replace doctors - it extends their reach. A radiologist in a major city can now effectively serve patients in rural communities through AI-assisted remote diagnosis.
And there's the skills angle. Microsoft announced plans to train 2.5 million Americans in AI skills in 2025 alone, focusing specifically on community colleges - the institutions that actually serve working-class students. 40% of small businesses already report some level of AI adoption, up 25% over the last three years, driven by affordable plug-and-play tools.
The pattern here isn't "billionaires get richer." It's "complex technology becoming accessible through better business models, infrastructure investment, and education." That's the same pattern we saw with smartphones, cloud computing, and the internet itself.
Does everyone benefit equally on day one? No. But the trajectory matters. And the trajectory for AI - especially in healthcare, agriculture, and automation - is clearly toward broader access, not concentration.
So what's the better story for AI?
It's not "we need to beat China." Most people don't wake up worried about geopolitical AI races - they wake up worried about rent and medical bills.
It's not "productivity gains for knowledge workers." That sounds like more money for people who already have money.
The better story is this: We can end preventable suffering.
Your grandmother doesn't have to die of Alzheimer's. Your kid doesn't have to develop the diabetes that runs in your family. The chronic conditions that consume 86% of healthcare spending aren't immutable facts of life - they're engineering problems we're finally equipped to solve.
AI is the tool that makes this possible within our lifetimes. Not because it's magic, but because biology is complex enough that no human researcher can hold all the variables at once. AI can.
If you're reading this and you work in AI, machine learning, or biotech - you have a job beyond shipping features.
The public is turning against our industry. Not because they're ignorant, but because we've failed to show them what we're actually building. We've let the narrative become "chatbots and layoffs" when the real work is "drug discovery and disease prevention."
That's on us. And it's our job to fix it.
The story of AI curing what kills us isn't hype. It's happening in labs right now - peer-reviewed, NIH-funded, Phase-II-trial-happening. Those of us who see what's being built have a responsibility to tell that story. We just need to actually tell it.
Hard, but not impossible.
Sources:
December 27, 2020 · 7 min read
Why the most frustrating paradox in machine learning applies to startups, careers, and almost everything worth building.
September 27, 2020 · 7 min read
Six months into SAP Eureka, I stopped seeing AI as hype and started seeing it as a tool that could actually solve problems. Here's how automating trade claims rewired my thinking.