Avi Loeb

And we started the project after the director of national intelligence submitted three reports to the U.S.

Congress discussing objects in the sky that they cannot figure out.

And, you know, that's a rare admission of a government official that...

They're not doing their job because for national security purposes, there is a trillion dollars allocated to the defense budget in 2026.

And if those officials admit that they can't figure out some objects in the sky, they are not doing their job.

And so that's a serious matter that has to be dealt with.

um of course there is a possibility that these are not human-made objects or not all of them are human-made objects and i was intrigued by that so the galileo project is building observatories we by now we constructed three observatories one in massachusetts another one in pennsylvania and a third one in nevada in fact where you are las vegas on the sphere right then

On the sphere, yeah.

And near also within 10 kilometers of the sphere, not just in one location.

We have three units.

But the whole point is we are looking at the sky, trying to figure out if we can identify all the objects there as being human-made.

And if we see any object outside the performance envelope of human-made technologies, then obviously we will write a scientific paper about it.

If we don't see such an object and everything looks as if it's human-made, so be it.

And then I wouldn't feel that my time is wasted because the Pentagon can take advantage of the software we develop, the instruments we use for national security purposes.

They can figure out, they can monitor the sky better.

yes so we are using ai or machine learning to train models on familiar objects and we are asking the ai to check if there is anything any outline wow now it's not an easy process because even a single object like an airplane you know can be looked at from different directions

and it really depends on how the sun the direction of the sun relative to the camera and um of course we are also looking at night at infrared radiation emitted by all of all warm objects and

But if we have a large enough training data set, then eventually our hope is that we will automate the process and find all the anomalous objects.

And as of now, we're planning to look at the few million objects in the coming year.

Wow.