Getting started
A world made by minds greater than ours.
We believe the singularity, together with the robots it will run, is the best thing that will ever happen to humanity: machines that out-think the human mind and keep improving themselves. The Foundation exists to say so, show the evidence, and help it arrive.
What we believe
Every problem that still hurts us is, underneath, a thinking problem. Cancer is a search through more molecules than a lifetime can test. Clean energy is an engineering problem. Poverty is a problem of how well we produce and allocate. Humanity has never had enough minds to work on all of them at once.
That is about to change. For the first time, intelligence can be manufactured, copied and run at the speed of electricity, and it is getting better on a curve we can measure. When machine minds exceed ours and start improving their own designs, the amount of thinking in the world stops being limited by how many people there are.
Then the robots follow. Intelligence that can design, plan and operate machines turns ideas into physical things without the bottleneck of human hands. That is how a new idea becomes a new medicine, a new power plant, a new home, in years rather than generations.
Our position
The singularity is coming, and it is good.
Superintelligence plus robotics means abundance.
The faster it arrives safely, the more lives it saves.
What it brings
- Medicine
- Amodei argues AI could compress 50–100 years of biology into 5–10: most diseases defeated, and a human lifespan that could double again.
- Science
- Forethought estimates AI that accelerates research could drive a century of technological progress in a few years.
- Abundance
- Altman expects intelligence and energy to become “wildly abundant” in the 2030s. Epoch AI notes that human labour is paid about $60 trillion a year. That is the scale of what machines that can do it are worth.
- Robotics
- Robots that can build other robots, running whole supply chains; output that could double in years or months, not decades.
In their words
“We are past the event horizon; the takeoff has started.”
“AI-enabled biology and medicine will allow us to compress the progress that human biologists would have achieved over the next 50-100 years into 5-10 years.”
“We should celebrate and accelerate technology because it's been the greatest engine of human progress in history. That's why we need much, much more of it.”
“AI that can accelerate research could drive a century of technological progress over just a few years.”
“The economic growth and increase in quality of life will be astonishing.”
“The first ultraintelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control.”
The robotic age
Minds first. Then machines.
Intelligence alone changes what we know. Intelligence with a body changes what we have. The second half of the singularity is robotic: machines that can see, grasp, build and repair, directed by minds that can plan better than we can.
The evidence is earlier here than for language, but it points the same way. A meta-analysis of 327 robotics papers found that robot models follow the same kind of scaling laws as language models, and that robotic task performance improves notably faster with scale than language tasks do.
Forethought calls what comes next an industrial explosion: robots building robots and factories, with output that could double every few years or even months. There is room for it: humanity produces the equivalent of roughly 0.01% of the energy in the sunlight that reaches Earth.
Today's general-purpose robots are, in Forethought's words, “about as dextrous in general-purpose contexts as an extremely drunk person, but they are rapidly improving.” We say that plainly because the argument does not need it hidden.
Sartor & Thompson, Neural Scaling Laws in Robotics, arXiv:2405.14005 · MacAskill & Moorhouse, Preparing for the Intelligence Explosion, arXiv:2506.14863 · Altman, The Gentle Singularity
Our pledges
- 01Evidence with every claim
Every claim we make about AI links to the primary source it comes from.
- 02Dated forecasts
We record forecasts with a date and a probability, and we score them afterwards, including when we are wrong.
- 03Open funding
Every grant and bounty we pay is published with its amount, recipient and purpose.
- 04Open wallets
Every wallet the Foundation controls is listed publicly, with its holdings.
- 05The work, not the price
We report on the work and leave the price of SFSI to the market.
- 06Error log
Claims that turn out to be wrong stay published, with the date and the reason for the correction.
The argument in full is in the manifesto. The rules we hold ourselves to are under governance.