Below are real cases of AI benefiting humanity, which are markers on the Road to abundance. A start, not a finish line. We are on that journey, which will require time and effort to improve everyone’s quality of life, including creating better quality, lower-cost healthcare, and the following examples are some of the markers along that path.
A marker on the road to abundance is better.
Parallel Bio organoids
evidence for this kind of research and development, which will benefit humanity, is described by Richard Socher on Moonshots with Peter Diamandis.
Richard Socher, on Moonshots with Peter H. Diamandis, recorded October 2, 2026: Parallel Bio builds tiny organoids on a petri dish. “They had FDA approval to skip animal trials. Turns out we cured most diseases in mice. It’s not that helpful.” He says they use pluripotent stem cells to grow tiny lymph-node organoids, and that those tests were more predictive of how drugs behave in human bodies than mouse work. He ties that to immunotherapy, the immune system attacking a cancer instead of heavy chemotherapy. Moonshots episode 299
Source: Moonshots, episode 299, about 17 minutes in.
- Richard Socher: “multiple different diseases will get cured in the next 12 months… one gene that needs to be fixed.” He adds better battery material, and he says it is not one threshold where in three years everything is solved. More problems get solved over time.
Examples, not settled claims. Milestones on a long road that takes years, not a doorway.
- Richard Socher: “Now you have the new age of companies. They have five to 10 different compounds in stage three trials after two or three years.” He puts the natural delays for scaled use at half a decade to a decade, and says chemistry and physics will be even faster.
- Richard Socher: tokamaks are already balancing plasma, a hard control problem where AI is being used. He treats fusion as still unsolved. This is a step underway.
- Dave Blundin, with Socher agreeing: a scientific superintelligence running a million square feet of robotic space, toward a thousand times the rate of discovery. Socher names Lila, and Periodic Labs on the physics and chemistry side.
- Richard Socher: “Mark Zuckerberg Foundation started a virtual cell. Everyone’s doing it.” Companies such as Tahoe Therapeutics are running massive perturbation studies. He says you cannot measure every protein in a cell without destroying it, so the virtual cell is a middle layer.
- Peter Diamandis: Brain-IT “reconstructs the image a person is looking at while inside a functional MRI machine.” Alexander Wissner-Gross: full-bandwidth brain interfaces will get there, and you have to start somewhere.
- Peter Diamandis: an executive order to cut invasive mosquitoes in Washington, D.C. “by at least 90%” and ticks “by at least 50% by 2028,” prioritizing sterile insect techniques, safe genetic modifications, and beneficial bacteria over conventional pesticides.
- Richard Socher: “This year, something major shifted, and that is AI can now code.” In weak forms that loop is already in use. He says stronger forms are close, not here, and that the strongest systems will take several decades.
- Alexander Wissner-Gross: “For decades, we’ve been scared of our shadow.” He says that fear cost decades of progress in engineering the physical world and the biological world.
Source: Moonshots, episode 299, recorded October 2, 2026.
Health
No story here yet. One will be added when it checks out.
Medical
These cases are evidence we are already on the road to abundance, not a sign the road is only starting.
- Joshua Resnikoff, chief executive of Sunstone Health. His son Shiloh went years without an answer. Resnikoff then built a company that uses AI to flag other children sooner, on the order of weeks instead of years. That is not a treatment he found for Shiloh. Source: EOFire, “From 7 Years to 12 Weeks,” with Joshua Resnikoff. eofire.com/podcast/joshuaresnikoff
- Matthew Might, a computer scientist, not a company chief executive. An AI reader of the medical literature pointed to a compound he used for his son Bertrand, who had NGLY1 deficiency. A later use of a related tool, mediKanren, pointed to ketamine for a different rare condition, ADNP syndrome, and that lead moved toward a clinical trial. Sources: Rare Disease Advisor, and Medical Xpress, May 2020. Rare Disease Advisor. Medical Xpress
- Mayo Clinic, May 2026. Doctors used an AI tool, Biomedical Data Translator, to match a newborn, Jorie Kraus, who has an ultra-rare chromosome 10 deletion, to a drug already on the shelf. Source: STAT, May 19, 2026. STAT
More cases belong in this category, or in new ones, as they check out.
Education
AI used as a personal tutor. The point is individual attention for people who have been held back from knowledge, including minorities who were denied an education, so they can become what they want to be. Handing someone a computer is not that. A case that checks out belongs here as evidence of that same road, already underway.
Examples, not settled claims. Each is a measured learning gain, not proof that a personal tutor is already in every classroom.
– Rori, Rising Academies, Ghana. Grades 3 through 8 who used the WhatsApp math tutor for two short study halls a week, on top of normal class, grew more in math than classmates who did not, about 5 points versus about 2. The authors call that roughly an extra year of learning, at about $5 a student. Year 1 of the trial. Source: arXiv 2402.09809, 2024.
– Tutor CoPilot, Stanford with FEV Tutor. It suggests wording to a human math tutor. Students whose tutors were offered it were 4 points more likely to pass the session exit ticket, 62 percent to 66 percent. Students with lower-rated tutors gained about 9 points. About $20 per tutor per year. Source: arXiv 2410.03017, 2024.
– World Bank Copilot pilot, Benin City, Nigeria. Six weeks of teacher-guided after-school English with Microsoft Copilot in nine public secondary schools. End-test gains were about a quarter to a third of a standard deviation, depending on the score. Source: World Bank Education Working Paper No. 18, 2025.
– Harvard physics tutor, Physical Sciences 2. A custom tutor, head-to-head against the class, two lessons. Scores were higher with the tutor, in less time, about 49 minutes versus about 60 in class. Source: Scientific Reports, 2025.
– Mindspark centers, Delhi. This is computer-assisted adaptive learning, not a generative tutor. For 619 students, mostly grades 6 through 9, a center with adaptive software plus small-group instruction raised math and Hindi over about 4.5 months. A teaching assistant was in the room. Source: J-PAL evaluation, 2015–2016.
Entrepreneurial
Instances of AI helping people start or grow a business. A case that checks out belongs here as evidence of that same road, already underway.
No story here yet. One will be added when it checks out.
P(doom) is zero means that the probability of humanity being wiped out is zero, per Richard Socher, founder and chief executive. Moonshots episode 299 P means probability. Doom means humanity being wiped out.
Richard Socher, founder and chief executive of You.com and co-founder and chief executive of Recursive, on Moonshots with Peter H. Diamandis, recorded October 2, 2026: there is no realistic scenario where AI wipes out all of humanity, and “P(doom) is 0.” He still said large harm is possible. His “maybe 100 million” was a discussion point, not a forecast. His reason is that as intelligence gets cheap, defenders can find a finite set of attacks and inoculate against them. Peter hosts and presses him, and on that episode Peter says he is a seed investor in Recursive. This is Socher’s judgment, not a measured probability.
The written defense argument is his essay AI Safety Paradox, May 18, 2025. That essay does not use the sentence “P(doom) is 0.” Source: Moonshots, episode 299.
Build toward. Don’t destroy.