May 2nd 2025
Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.
00:00 - Tariq Malik (Host)
Coming up on this Week in Space. The Trump budget is out and it doesn't look good for NASA. Planet 9 does exist and we're going to find out what the deal is with AI in space from Dr David Bell at USRA himself. So tune in and find out.
00:15 - Leo Laporte (Announcement)
Podcasts you love.
00:17
From people you trust.
00:20
This is TWiT.
00:23 - Rod Pyle (Host)
This is this Week in Space, episode number 159, recorded on May 2nd 2025. Ai in Space. Hello and welcome to yet another episode of this Week in Space, the AI in Space edition. You're going to want to stick around for this one. I'm Rod Pyle, Editor-in-Chief Bad Aster Magazine. That's not why you're going to want to stick around, and particularly not because I'm joined by my fellow non-mathematician, Tarek Malik, editor-in-chief of Spacecom.
00:49 - Tariq Malik (Host)
Hello, my friend, hello Rod, how are you doing today? Happy Friday, happy podcast day.
00:54 - Rod Pyle (Host)
I'm okay, and happy day after May Day, that's right.
00:57
In a few minutes we're going to be joined by Dr David Bell, who's the director of the USRA Research Institute for Advanced Computer Science which is as cool as it sounds and is the USRA program manager for the Quantum Artificial Intelligence Lab, collaboration between USRA, google and NASA's Ames Research Center. And if that doesn't interest you, then we're in the wrong room. Before we start, don't forget, please, to do us a solid and make sure to like, subscribe and the other cool podcast things for us, because we're counting on you. And now, yes, a space joke from Anonymous.
01:34 - Tariq Malik (Host)
Anonymous. Ooh, I like that. That's new hey.
01:41 - Rod Pyle (Host)
Tarek. Yes, Rod. Why don't AIs get lonely in space? I don't know why. Why don't they get lonely? Because they always have their neural networks for company.
01:51 - Tariq Malik (Host)
Oh, that's good Wah wah, wah.
01:57 - Rod Pyle (Host)
Now I've heard that some people want to stick us into a neural network when it's joke time in this show, and won't that be disappointing us into a neural network when it's joke time in this show, and won't that be disappointing? But you can help send your best, worst or most indifferent space joke to us at twisted twittv and, uh, if you're not ashamed of it, you don't have to be anonymous.
02:15 - Tariq Malik (Host)
Uh, by the way, because you've got robot to do, right, like a lot to do. No, I got it.
02:22 - Rod Pyle (Host)
I just wasn't reacting, cause I I was. I was feeling a little nauseated, by the way, as I've pointed out before, but this is particular to this episode. If you happen to be in florida in june, tarik and I will both be at the international space development conference that the nss puts on every year, and I'll be hosting a plenary session on ai in space with three very prominent space visionaries on Saturday, which is going to be a lot of fun, and I picked three of the smartest people I could find. So if I can get my laboratory retriever-like brain up to speed, I'll be able to ask them some challenging questions.
02:58 - Tariq Malik (Host)
So my invitation got lost in the mail. Is that For that?
03:02 - Rod Pyle (Host)
panel. The mail is that for that panel. Hey, you're on a plenty of panels and our sunday lunch where you'll be getting your space pioneer award. So I think that's plenty. All right, let's do some headlines. Oh, oh, wait before we do sorry headline headline newsline news.
03:22 - Tariq Malik (Host)
Headline news. I got it, I know it now.
03:26 - Rod Pyle (Host)
Wait, wait. What are we doing before? Okay, sorry, I meant to say if you're interested in the ISDC, you can get more information, including the speaker's roster, at isdcnssorg there you go. Well worth looking at. If you're there, we'd love to see you and you can come say hello to me and insult Tarek while you're there.
03:45 - Tariq Malik (Host)
Okay, my mom will be there. Don't insult me.
03:49 - Rod Pyle (Host)
I'll be, I'll be good. I promise it's not easy, but I'll be good. So headline one list Trump administration proposes slashing NASA's budget by 24%.
04:01 - Tariq Malik (Host)
It's the biggest NASA budget cut in like modern history compare, I mean probably since the Space Race, I'd say yeah, yeah, and it's very skewed, and this is this is very fresh, by the way, just so that everyone knows this came out literally a few hours before we started recording the podcast.
04:18 - Rod Pyle (Host)
So so it's been talked about up till now, but this came out in the skinny budget, right?
04:22 - Tariq Malik (Host)
yeah, this came out in the skinny budget. I haven't gone through all the details, so what we're going to talk about is very top line, but it is not looking great. It could have been a lot worse. We talked on the show in the past about a potential 50% budget cuts to science, etc. But this is pretty strong in its own. This is a skinny budget for fiscal year 2026 from the Trump administration that would cut NASA funding by $6 billion. So that means that from 2025, enacted levels $24.8 billion. They're saying no, going back down to $18.8 billion. So that's where that 24% comes from and it ends a number of programs. It means that the Gateway space station around the moon probably going to go Artemis, done after Artemis 3, which includes.
05:17 - Rod Pyle (Host)
Well, artemis, as we know it today, I'm saying SLS, sls.
05:21 - Tariq Malik (Host)
The SLS is what. I'm talking about SLS done after Artemis 3. And then I guess Artemis would evolve after that.
05:27 - Rod Pyle (Host)
Excuse me one second. I also read and I'm not sure if this is accurate or not was that SLS and Orion would go away after Artemis 3. But Orion could be launched on a variety of other spacecraft, can't it?
05:41 - Tariq Malik (Host)
They could potentially adapt it for that right. And so I mean, I think that there was talk about Atlas 5, and maybe when the site gets phased out, you could look at Vulcan, that sort of thing. They've got options with that one because they were able to save it from when it was part of Constellation, etc. So I think that that might be up for discussion. Mars sample to return, another one that could be on the cutting room floor in at least its current form, because of the frustrations and the budget issues right now.
06:14 - Rod Pyle (Host)
Okay, what about the Nancy Roman Space Telescope, which was rumored to be on the chopping block? And is finished and ready to fly, by the way.
06:21 - Tariq Malik (Host)
That's what I have to look into right now because, as I said, like I was looking at the human spaceflight you know I'm a rocket person, so I look at the human spaceflight stuff the most and human space exploration overall actually received extra money 650 million more but it has these other things that look like they're on the chopping block. But what I can say because the Nancy Grace Roman telescope is, you know, a purely science application and, as you said, already built etc. We talked with Dr John Grunsfeld on our last episode about that that the planetary society, you know, which has been decrying the potential science cuts, is, you know, really trying to make it known that the science budget overall is getting gutted in this budget. So they're saying, you know it recklessly slashes up to 47% with widespread terminations of functional emissions that are worth billions of dollars. This is from their statement today and so they're really pushing to get Congress involved to really, I guess, step up and say, look, we need to preserve science. And so if the Roman stuff does carry through, it would be extremely disappointing, because we've seen that from a Trump administration in the past to just cancel science stuff where it's built already and some of that was able to be restored back in the first administration when the skinny budget came out and they proposed canceling five different, I believe at the time, earth science missions different, I believe, at the time, earth science missions. Once Jim Bridenstine came in, they were able to restore at least a good number of those, if not all of them. They did not get rid of them. Some of them were active missions, some of them were yet to fly.
08:17
We're still waiting for Jared Isaacman to be confirmed in the Senate as NASA administrator. He has said that he does not feel that these drastic science budget cuts are a good idea or make sense. He's on the record saying that to the Senate subcommittee but he's not administrator yet so he doesn't have a say and so there is still a lot in play as to like what's going to happen with these cuts. But it is a really deep cut Again very, very fresh. I haven't gone into the line items yet from the skinny budget, so I really want to be able to do that and maybe next episode we can come back and say, look, this is what I found. This is who's squawking, who's saying it's a good idea, who's saying it's a bad idea. And we can, maybe we'll do the whole episode on that Rod.
09:00 - Rod Pyle (Host)
I don't know We'll have to see is is there a particular person we can think of that has let's see an operational space capsule? Who might want orion shot down?
09:12 - Tariq Malik (Host)
well, I mean think of anybody.
09:13
I, I, my name, my, my name, my, my, my, my mind is drawing a blank rod, as you all know. No, the, the gutting of the SLS and the Orion very much makes sense if it's something that comes from a billionaire owner of another space program that is trying to get customers for their ginormous rocket. I'm not saying that that's the case, but we know that Elon Musk has been very much involved in both the budgeting of the current administration as well as a lot of insight into the space agenda. We'll see how much of that really sunk in through whatever he advised the administration to do and how much that involvement was. I think it's still kind of up in the air how much of it came from that.
10:02 - Rod Pyle (Host)
Well, and I think we all want to see Starship succeed. If it turns out to be the right architecture, I think there's still a question mark there.
10:11 - Tariq Malik (Host)
But it means you're not going to get the exploration upper stage for SLS.
10:14 - Rod Pyle (Host)
It means we're not going to get to the moon by 2030, frankly.
10:17 - Tariq Malik (Host)
That's what it seems like. That's my opinion.
10:27 - Rod Pyle (Host)
But when you have somewhere between 16 and 24 retanking flights for one lunar foray that have to take place for a spacecraft that, let's see, it's, uh, the middle of 2025 and it still has not had a successful full orbital flight, yeah, gosh, that's kind of a problem. Which brings us to, yes, a couple of kind of surprising press releases from nasa over the last weeks. The first one, I literally I pulled it up and I kind of gulped and choked a little bit. And then the second one, I was a little less surprised and you know, I understand and you're going to explain to us why I'm saying this, I understand why NASA felt they had to do this. Maybe it's kind of specific to Janet Petro, maybe it's kind of specific to janet petro, who's the interim administrator, I don't know, but they were fawning over the executive branch should we?
11:12 - Tariq Malik (Host)
should we say what it says? Nasa soars to new heights in first hundred days of the trump. Is that what we're talking about?
11:18 - Rod Pyle (Host)
yeah, yeah, that one and uh, president trump's fiscal year 26 budget revitalizes human space exploration they forgot the and kills science, especially climate science part.
11:30 - Tariq Malik (Host)
Yeah, that to me feels like putting a lot of lipstick on a pig right? So they have to. They have to spin their budget even though it has a 24% cut across the board as as much as they as they can. And so what they are focusing on is like hey, we've increased our commitment for human space exploration by giving them more money to do more things, but we're not going to talk about the fact that we're shutting down, or would shut down, a whole SLS rocket launch system, which, by the way, was signed into law. Right, congress ordered NASA to make that rocket.
12:05
So they have to change the law to cancel that program.
12:08 - Rod Pyle (Host)
And that would be the moon rocket that we have. That actually works right, yeah, right Okay.
12:14 - Tariq Malik (Host)
And so it's a lot of human, spaceflight-focused stuff about science, technology stuff, all about putting an American on Mars, not China getting to the moon before China. That's what you're seeing in this list here, and so I understand why they put that out there, because there was always a press release about the budget and they have to try to paint as positive a picture that they can. But there is a lot of politics wrapped up in this one this year because of just the state of the government and how I guess how divisive the administration has been, especially with the cutting of budgets and staff at agencies, and then just how much the rhetoric has gotten into nasa. We saw it with the moon landings earlier this year where, uh, they're saying, you know, america first on the moon and and whatnot, uh, and and it's, it's just, it's just stuff that we have to watch out for, uh, so that you're you know you're not taking it all at face value, that you see the context behind a lot of it.
13:21 - Rod Pyle (Host)
So okay, let's, let's bang through a few more of these so we can get to the show. Show Planet 9, Planet X, whatever you want to call it. It's a cool story and we know that for many of us, even those of us who know Mike Brown at Caltech Planet 9 kind of still is Pluto. But that's not what we're talking about. We're talking about Planet X here.
13:44 - Tariq Malik (Host)
That's right, that's not what we're talking about.
13:45
We're talking about Planet X here, that's right. That's right. So an as yet unseen planet out there beyond Neptune, where we think that it's having gravitational effects on the Kuiper Belt objects that are out there, but it is very distant, it's very dark, it's hard to see and they've been trying out to hunt for it. Now, the reason I picked this one, it's just because it's I mean, it's fun to talk about Planet X or Planet 9 or whatever you want to call it a hidden planet in this day and age, in our solar system.
14:12 - Rod Pyle (Host)
Excuse me one second. Can I just back up a step? The reason this isn't as wacky as it may sound to some is that by observing planetary orbits and you can look at them and say, hey, there's something else affecting these out there that makes them look the way they do. So it's not that strong grounding. It is how we found Neptune.
14:33 - Tariq Malik (Host)
Yeah, yeah, it's exactly how we found Neptune and Planet Nine as it will.
14:37
It was really named as such by Michael Brown and Constantine Badigan at Caltech back in 2016, from what they've seen and it's kind of a dig at the Pluto lovers right Just call it Plano 9, because in 2006, right was when they demoted Pluto.
14:57
And the reason that it's in the news again is because scientists have gone back and looked at old data from the 80s with the infrared astronomy satellite I think we're going to talk about IRAS later in the episode too and realized that they found what looks like evidence of Planet Nine in some of the object motions that were observed by both IRAS as well as Japan's Akari satellite, which was a launch between 2006 and 2011. And that would be more evidence from the archives. That lends itself to this actually being a planet that's out there and maybe a bit closer to being able to see it. So you know it goes back 40 years. An object that is about 700 astronomical units out from the Sun. We are one astronomical unit out, so that puts it in perspective 700 times as far as the Earth is from the Sun.
15:59
Very exciting stuff.
16:01 - Rod Pyle (Host)
And let's do a speed round. We've got concerns about the Soviet Venus lander, we've got concerns about the Psyche mission and I have concerns about Gen Z and millennials. But you first.
16:13 - Tariq Malik (Host)
Yeah, yeah. Well, let's look at Psyche really quickly. Nasa's Psyche spacecraft, which and the iron astro Pardon me an iron asteroid has got problems. It's lost some fuel pressure. They're trying to work through it. Hopefully they'll be able to resurrect it. I forgot to put the link in here, so I'm sorry we can't show Psyche. They're hopefully trying to be able to recover it. We hope so too, because I really like that mission and everyone got tattoos of it. Oh, you've already got the link in there. Look at that. That's great. Wait, you have a tattoo. No, a lot of the mission team people got tattoos.
16:55
I thought we were going to get personal here, no, they got the Psyche asteroid that it's going to. They got renderings of some places on their persons, if you will. They all got to pick where it was. Jim Bell didn't tell me where he got his.
17:08 - Rod Pyle (Host)
I'm going to be good.
17:12 - Tariq Malik (Host)
And then we have the Soviet lander. We have new images of the Venus lander that is falling out of space, the Cosmos 42 Venus probe, which people are tracking over time. It was launched into space back in 72, but it didn't actually make it to Venus, obviously, because it's coming back down to Earth and folks have been tracking it. Marco Lengbrick at SatTrakCam in the Netherlands, among others, have been watching it just get lower and lower, and we're seeing imagery now from their telescopes as they track it to see like, where is it going to fall? They think it's going to reenter around May 10th.
17:53
As we're recording this, it is May 2nd and so we're trying to see. You know, where is it going to fall? Is it going to be over a populated area? Most likely not. Most of our planet is covered in ocean, but it's a reminder that there's a lot of stuff up there that didn't end up where it's supposed to be and it's all going to come back down. Like the old adage, what goes up is going to come down, so this is just really exciting. It's another example of a little bit of vintage history that you can track and see over time.
18:20 - Rod Pyle (Host)
Cool. And finally and I don't have the source for this, I'm afraid because it was a radio item this morning when I was prepping the show that said, if I've got this right, just over half of Gen Z and millennials believe in alien coverups, that area 51 has things that we don't know about and that they're real and in some cases they walk among us, and so on and so forth. Now we recently saw a poll from was it Ipsos or I forget which organization it was in like 2019, 2020. It was the first, or maybe it was 22, the first global poll of space attitudes. They did address this question and the beliefs tend to be high in younger people globally.
19:07
But the US is particularly rife with folks that believe in aliens, and not just aliens, because, I mean, I believe in the idea of aliens because it's a big universe there's a lot of planets out there, right, but this is in particular. They're here, they've been here, the machines are here and in some cases, they walk among us. Particular, they're here, they've been here, the machines are here and in some cases they walk among us, and I can't say this gives me a lot of warm fuzzies.
19:32 - Tariq Malik (Host)
But you know, it's just one of those things. Well, I mean, it's like there's those people that don't think we're on the moon, right, and they maintain to this day that it's impossible. You know, clearly, because we can't go back, apparently. Oh my god.
19:43 - Rod Pyle (Host)
So, but well, let's not forget that ns31 was fake too, because they're lacquered with hairspray.
19:51 - Tariq Malik (Host)
Hair didn't fly free, I mean yeah, yeah, there's that oh my gosh you know, by the way, katie perry was also in the news this week, uh lashing back at the criticism uh of everyone, uh decrying and saying how bad that flight was and everything. So that's a mission that is still in the headlines, going on while she's on tour.
20:11 - Rod Pyle (Host)
Hey, peace, love and Bobby Sherman, everybody All right. We will be right back in just a few moments with Dr David Bell. Stay with us. And we are back with Dr David Bell, who is the director of the USRA Research Institute for Advanced Computer Science, which sounds like about the coolest place to be working in the world right now, and also working with and correct me if I get this wrong the Quantum Artificial Intelligence Lab, collaboration between the USRA, google and NASA Ames Research Center. Did I have that right? Mm-hmm, which sounds pretty amazing, so I hope we can talk about that at some point. And you also focus on collaborations between universities, industry and NASA in a range of domains, which includes machine learning, autonomous systems, nanotechnology and biotechnology, but also quantum computing, which we're hearingotechnology and biotechnology, but also quantum computing, which we're hearing more and more about, and, if we have time, I'd love to learn more about that. But how long have you been with this organization?
21:18 - Dr. David Bell (Guest)
I've been with USRA 20 plus years. Before that, I worked for about 10 years at the Xerox Palo Alto Research Center.
21:27 - Rod Pyle (Host)
So you've kind of seen it all since it's begun to emerge, haven't you?
21:31 - Dr. David Bell (Guest)
Yeah, absolutely. The group I was in even before coming to USRA was the scientific and engineering reasoning area of Xerox PARC, and so that was a group of artificial intelligence computer scientists looking at the application of AI for science and engineering domains.
21:51 - Rod Pyle (Host)
So, seeing that you've been involved with this as long as you have, was the emergence of the large language model architecture a big shift for you folks, or was it just another step in what you've been doing all along?
22:06 - Dr. David Bell (Guest)
No, I think it's definitely represents a pivotal moment. My favorite book is Diffusion of Innovations by Everett Rogers, where he describes these S-curves of technology evolution, adoption and saturation into different markets or, around the world, different walks of life. And you know, you start off with slow tail, low tails of adoption as technology evolves and then you have this phase with exponential infusion really broadly and then gradual saturation as the technology gets out there. And you know, if you think about historically, some big technologies like the internet right, started in I think, the 60s with ARPANET, but then in the 1989, 1991 timeframe, tim Berners-Lee and others created the Hypertext Transfer Protocol, http, the World Wide Web, chrome browsers, and that really caused an exponential explosion for more nodes on the Internet and really infusion of that into all walks of life. And so with artificial neural networks, I think the first artificial neural network was like 1959. Right. And then Xerox, palo Alto Research Center we were doing AI there.
23:28
I had AI back in the 80s in in college and but in you know, the 2017 invention by Google of the transformer architecture.
23:45
And then 2018, openai created GPT, the generative pre-trained transformer and then later released ChatGPT I think to the amazement of us all and then later released chat, gpt, I think to the amazement of us all.
24:00
And then there's just been billions of dollars of investment and what we're seeing is the size of these large language models is growing exponentially. It started with 10s of millions, then hundreds of millions of parameters, then now over a trillion parameters, and the performance of these models and all these different benchmarks math, chemistry, sats, you know has been approaching human performance level and then exceeding and then approaching expert level in different areas, and it's continuing to evolve. So it's definitely a pivotal moment, I think, in the history of AI and we all have to adapt to it and we've adapted to it. In addition to the quantum AI lab, which we think is one of the future technologies that will help increase the energy efficiency and performance of training and inference of artificial intelligence, we've also last year created a generative artificial intelligence lab for science and engineering, where we're we're working on something which I think is a new, a second wave of large foundation models like the LLMs.
25:15 - Tariq Malik (Host)
Well, David, you know it sounds like you're very much working in like the future. Like right now is what it sounds like you're very much working in like the future like right now is what it sounds like you're doing there with the USRA and whatnot. But I'm very curious about your path just to this moment, because you mentioned about being in college in the 80s and I was really struck because when I meet a lot of scientists or engineers or even astronauts, they studied one thing at this place and one thing at that place and I was really struck that you had a pretty through line at Cornell with engineering all the way up to your PhD there. But I'm curious, what puts you on that path in the beginning? Is just that computer science, that engineering, a bug that you picked up when you were a kid, or was the space part of that there and you were trying to find a way into that and computers were that method? I'm curious what that path was like to get you to the point that you are now at USRA.
26:16 - Dr. David Bell (Guest)
So it's a fun path. I'm from a family of engineers and when I went to Cornell I kept wanting to and I did internships in engineering and I kept wanting to find ways to have a bigger influence on society. And so I started just doing engineering of a subsystem. And then I said, well, let's do computer aided engineering of tools that could help engineers. So then you do one thing and it helps thousands of people. And then at the time there was robotic automation in manufacturing. That was the first wave of automation. And then there was robotic automation in manufacturing. That that started. You know, that was the first wave of automation. And then, then we there was a group of us that started looking at, well, how can we bring automation into research and development? And so there was the national science foundation funded design theory and methodology program, which I I was part of, and we were looking at how we could bring information technology into computer science, into the research and development program.
27:27
In high school we had a Wayne computer with a cassette tape and you know I wrote like the biggest programs they ever saw. I fixed their punch card readers for surveys and throughout the software to analyze, calculate the statistics. And then I got the opportunity to go to Xerox PARC my advisor when I was in my PhD program. He took a sabbatical leave to UC Berkeley and he said, well, why don't you come out west? And I said, okay, you know, it's like I had connections through Xerox and Xerox Park and they said, okay, great, we'll do an internship. That internship ended up being for not just a summer but a whole year and then when I graduated, they said David, why don't you just come work here, you know? So I went back. That was my job search.
28:16 - Tariq Malik (Host)
Well, I totally understand that, as someone who had found both of his reporting jobs as interns, like I totally get that. But it sounds you kind of skipped over it that you were working on computers in high school, because I mean, up until that point when you mentioned that, it sounded as if the whole engineering and computer science that you've been working on was an organic discovery that came through your engineering work at Cornell. But no, you were like into this in high school. That sounds cool.
28:44 - Dr. David Bell (Guest)
That was like 80, 81 dates me, but yeah, and then you know, really interesting thing you know for USRA and our history is and this surprises people when I tell them but in 1983, ronald Reagan was the President of the United States. Every year, by law, the President of the United States sends a space and aeronautics report of the President to the US Congress. In that report that he transmitted in 1983, there's a paragraph that says industry is making rapid advances in supercomputer technology, human computer interfaces and artificial intelligence. This is important to the space industry and, as a result, we're creating a new computer science initiative at NASA and making a major commitment to work with the University Space Research Association to establish an independent research institute for computer science operated by USRA at NASA's Ames Research Center, which is in Silicon Valley.
29:55 - Tariq Malik (Host)
That is so cool.
29:56 - Dr. David Bell (Guest)
And so our institute's been doing high-performance computing, human-computer interfaces and artificial intelligence ever since, and you'd be really surprised with the infusion of these that's been happening into NASA missions starting in the 1980s, where AI was making astronomical discoveries already in the 1980s.
30:16 - Tariq Malik (Host)
See, that's amazing, because the only thing I knew about artificial intelligence and computer science in the 1980s See, that's amazing, because the only thing I knew about artificial intelligence and computer science in the 1980s was what I learned from Tron when it came out in 82.
30:24 - Rod Pyle (Host)
You know, well, I want to give you a big slap on the back for sticking with it. Now, I'm a little older than you, probably more older than you. So my first computer class was in 1968 at the Museum of Science and Industry in Los Angeles on a system called CyberKnack, and we did a lot of carrying of stacks of punch cards around, which were really cool until you dropped them.
30:47
Hey these aren't numbered. Oh, I got to start over. We're going to run to a quick break and then we'll be right back, so everybody stand by. So, before we get too deep into this, could you help us understand? Well, maybe you're just helping me understand. I suspect a lot of our listeners probably already know, but what are the key differences between the AI we're seeing now general AI, which is the next step, and agentic AI, which, as I understand it, is kind of lateral to that in a way?
31:19 - Dr. David Bell (Guest)
Yeah. So a lot of the types of space related activities that that AI is being applied to in a way are the same today as they were back in the eighties. So planning and scheduling for missions, analyzing data from, you know, space telescopes, earth observing satellites, robotic locomotion, natural language processing, all of these things were being done even back then and solutions were being made. So in the 80s one of our scientists, peter Cheeseman. He led a project creating a software called AutoClass, which was an unsupervised machine learning approach algorithm that was used on a first-of-its-kind NASA Space Telescope mission, iras. That was the first to survey the night at infrared wavelengths. It made an astronomical. This auto class was automatically classifying the data coming back from that mission and it discovered a new class of infrared stars. I talked to Peter about it. He said there was some. You guys will enjoy this. There was some debate at the time.
32:38
Who gets credit for the discovery? Is it the software? You know? Does Autoclass get credit? Is it the computer scientists and scientists who wrote the software? But in the end the computer scientists and an astronomer published a paper in an astronomical journal and you know that's who gets credit for the discovery. But you know, kind of fun story.
33:01
So more recently. So that was Autoclass and 10 years after it was invented, lots of patents cited from Microsoft and other things. It was used in other fields. National Institute of Health projects used it. But you know, the algorithms keep evolving.
33:22
And so more recently, just a few years ago, one of our scientists led the development of a software, again for a first of its kind NASA mission, the Kepler mission, which was the first planet hunting mission of NASA. So you get the data back from Kepler and you want to analyze it to validate whether these are planets, and so the software ExoMiter was used to validate more than 300 new exoplanets. And it wasn't auto class anymore, it was a deep machine learning approach, which is one of the more recent inventions, and it was explainable. This is another thing that AI is trying to get away from just being a black box to being able to explain how it came up with its answers Same types of problems that you're trying to do automatically analyzing scientific data from space telescopes. It started with auto class. Now it's deep and explainable. Now your question was also about kind of agentic AI and you know agents, basically agents, If you think about you know your travel agent or something.
34:42 - Tariq Malik (Host)
they do work for you, so not like Agent Smith from the Matrix. Right, we're talking about a helpful agent, Helpful agents.
34:47 - Dr. David Bell (Guest)
Yeah, as long as you you know, train them well and code them well. So, agents we've also been doing agents for a long time. So we had a project that was called mobile agents, where you're not coding them with procedural software like, if this, then that You're having them sense their surroundings at least the ones we did for mobile agents and not all agents in the current nomenclature of agentic AIs like this, but the ones we did were mobile agents where they sensed their surrounding, the information around them, and then, based on the sensors that they had and the information they gained, they acted, and you could have more than one of these agents sensing their surroundings acting, and then other agents sensing and surrounding and then working together collaboratively. And so we actually deployed this into production use for admission control and interfacing with the International Space Station 20 years ago, and in that case there was an orbital communications officer who had a job of synchronizing some files between the ground and the space station. It was a very tedious job, and so the mobile agents were originally used just to simulate what the humans were doing, and then, once it did that, then actually they said, well, hey, the humans would like to get out of the loop. It's not a very interesting job for them. They'd like to do other things if they could, and so they put the mobile agents in place to actually do the work that the human was previously doing. And then the the humans actually then went off and were able to do more interesting work.
36:48
Now, more recently, agents, you know, when you go to chat gpt or gemini or any of these other services, you ask it a question, you give it prompt and it gives you an answer, but it's not taking actions for you. And so a Gen Tech AI is just saying, hey, we're going to analyze what you're asking us to do and then we're going to take action. Maybe it's to filter your mail or help create a summary, but it's an action. Or a or help create a summary, or you know, but it's. It's a seek, it's an action or a sequence of actions that will take place after you know some input.
37:26 - Rod Pyle (Host)
But again, agent based AI has been around for 20, you know 20 years, and then in use in space missions, in production use for space missions for that long too so for a number of years I was writing annual reports for uh jet propulsion laboratory and one of one of the projects they were working on, excuse me, was ai in robotic space flight fly through failure. You know, decision making in the outer solar system, all that. So I was talking to the chief engineer about that and this is more recently, and I said so what's the use case for LLM based stuff in this versus other things you're working on? And his sense of it was you know that there's a little too much comparative guesswork going on as opposed to absolute determinations, and I mean I'm explaining this from the standpoint of the consummate layman. Right, he is not, I am. So is there a certain evolution that has to happen before we start allowing this to make mission critical decisions like you're talking about with agentic AI?
38:31 - Dr. David Bell (Guest)
So let me give you there's kind of two parts to what you asked there. One is about this anomaly flight control part of where AI can be used, and the other is how will LLMs be used in robotic or human space flight? So on the first part of that, again I'll give a nice historical example. Okay, so one of the projects that we worked on was to develop AI solutions for planning and scheduling of activities, and so with the Mars Exploration Rovers, spirit Opportunity you had to plan the work of those rovers every single day and there's lots of constraints. Opportunity you had to plan the work of those rovers every single day and there's lots of constraints. You have to balance the energy use with how much energy you have. There's certain instruments and motions that can occur simultaneously and some that have to happen in sequence or at different times, and so you have all those constraints encoded into this planning engine.
39:34
Europa was the software tool, and then this was used every single day to plan the work of those Mars rovers to do what they did up on Mars. That same software then got used in another mission called Deep Space One, which was the first it was the part of this new millennium program, the first space mission to use ion propulsion and many other technologies, and so an AI engine was created with. There were three AI tools that were launched on Deep Space, one as part of a remote agent, again agent-based AI. So for remote agent it had three software, three AI parts to it. One was the planning and scheduling engine, which was built off of Europa. There was an executive that would help execute the plans and schedules, and then there was an anomaly detection and recovery software called Livingston, and so they actually had remote agent control the spacecraft for two weeks when it was up in deep space, and then they simulated failures, different failures. So they simulated a thruster being stuck on, they simulated a sensor failing, they simulated a camera issue and they simulated a problem with an electrical unit. And so this Livingston diagnostic and recovery engine detected those simulated failures and then enabled the recovery of the spacecraft from those failures. And so that was done in the early 1990s, I think I forget the exact date of it.
41:34
And so that same Europa planning and scheduling engine has been used for Mars science laboratory, mars Phoenix lander, I think it's called. It's also being used for human space exploration. So on the space station, the activities, the daily activities of the astronauts have to be planned. They have the exercise bike. Who gets to use it at what time? You know, that's one of the things you can't have. Everyone try to use it at the same time, and so those daily plans are now created with the help of this planning.
42:21
The evolution of this Europa planning and scheduling engine, which has evolved into a tool called Spiffy, and then there's a playbook version of it which is on an iPad and it's kind of like, if you know, google Docs, where two people can be editing at the same time and you see the changes real time. This planning and scheduling tool now allows distributed collaborative planning as well. Where you might, you know you can. The AI engine is there observing the constraints, but humans can be human assisted, where you're dragging things around and you see those changes simultaneously. The second part of your question remind me the second part of your question.
43:01 - Rod Pyle (Host)
I'm so fascinated with what you're saying I'm not sure I remember. Well, I had one that kind of came out of that though, because let's go to a break real quick and then we'll come back with your question, because I can see you're bursting at the seams. We'll be right back, don't go anywhere.
43:20 - Dr. David Bell (Guest)
I hope I'm not bursting at the seams because We'll be right back, Don't go anywhere.
43:22 - Tariq Malik (Host)
I hope I'm not bursting at the seams because, man, I let that go too far. That's going to make a mess everywhere. So, no, no, you know, I was going to kind of ask about that role in human spaceflight because a lot of the discussions and the use cases that I mean I'm surprised that it's been used in one form or another since the 90s and earlier, as you've been outlining. But they're really on that like back end kind of really facilitating the science or the planning of everything. And when I think the layperson thinks of like AI in space, they think of a companion that helps astronauts get through the day. You know, Gertie in Moon Hal 9000 in 2001. And I know that there is a side of artificial intelligence and research that is happening and I'm very curious how the evolution that you mentioned of these systems has either progressed or still needs to progress to get to that state Pardon me, oh my gosh, I'm getting a horse, I'm getting all emotional but to get to that point where that agent for an astronaut can go do a task.
44:37
Hey, go find the status of the O2 scrubbers for me and change it if it needs to be changed that kind of thing.
44:43 - Dr. David Bell (Guest)
Right, great question. So you know that original charter back in 83 was on supercomputers, human-computer interaction and artificial intelligence, and actually at NASA's Zames Research Center there's three facilities that are co-located so you could have all three disciplines work together on solutions that took into account all of these three same things. When I joined REACS and became its director, 20 plus years ago now, we had a group that was doing natural language processing. So one of the key roles for AI with humans can be to support natural language interfaces for astronauts. And so one of the projects was called Clarissa, and five years before Siri was released we had Clarissa up on the space station that had a conversation with an astronaut and it was a procedure browser. So you start simple, you know, because normally you know you have one astronaut outside the doing a EVA and you have another one flipping through their procedure book, kind of saying, hey, step one, step two, step three. But here an astronaut could ask and say what's the next step? Please repeat that. So it was very limited at the time, but it was the beginning of the support for natural language interfaces for astronauts and I think with the large language models that's one of the key roles that will be helpful is to enable, especially voice-enabled or text, but voice-enabled interfaces for astronauts. We are doing some work.
46:34
Nasa Johnson Space Center leads a leads a human research program for NASA, leads the human research program for NASA, and one of the projects we have is taking all of the NASA life science data going back to Gemini and you know, from Gemini on forward and it's in a database. There's archivists who put you know from Gemini on forward and it's in a database. There's archivists who put you know metadata onto these records and you have. There's omics data too and there's some public version of it. There's a private version of with additional data, and we've been, you know, helping NASA just with basic access to this but building a large language model index to it so people can give prompts, ask it questions, and it got so good that one of the line element scientists in the human research program in creating one of these summaries said, wow, that's something I might have written myself. That's hot and I think you know it's coming. One of the other really interesting projects out of NASA, ames, is called Astro B, where up on the space station where you have microgravity, I've seen those yeah.
47:52
Yeah.
47:53
And and, uh, you know it can, it can fly, it can maneuver around the space station and uh, so you know, so you can imagine in the not too distant future where you're having voice commands and and free flying robots flying around and actually you know doing more than just you know display screens and taking pictures and things like that you could, but you can take pictures, you can do things like that too.
48:23
Yeah, we also, for for there were 20 years ago we had, we did simulation. There's a Mars scape at NASA, ames where with kind of simulated regolith, this you know dusty surface of the moon or Mars, and you can have robots going out there, you can have astronauts walking out there. And so we simulated where you'd have a rover, a ground-based vehicle that would you could follow the astronaut or do actions for the astronauts as they're out on a, you know, outside of the space. You know the habitat, right. So there's lots of research been going on for 20 years with these kinds of robotic assistants and you know, as AI keeps evolving, the capabilities of what they can do and how effective they are in doing that just keep growing.
49:23 - Rod Pyle (Host)
So Tarek brought up 2001 A Space Odyssey. You probably know where I'm going with this, and if I was, let's say, nasa had the lack of foresight to send me to the space station. I would not want to wake up some evening with Astro B hovering next to my face telling me I'm using too much oxygen and I need to somehow be shut down for a short period of time?
49:43
Wait, I'm not a machine. So I guess, as this stuff gets more intelligent, more independent in action, the question of guardrails must come up. How do you address that?
49:56 - Dr. David Bell (Guest)
That's a good question. So the kinds of actions that the AI so far have been allowed to do are limited. When those mobile agents could transfer files between the space station and ground mission control and there's not much trouble they can get into there. It wasn't mission-critical data, it was more emails for astronauts and things like this. Pictures, with the robotic assistance that we've tested with taking pictures. Go to this location, take a picture. Tested with taking pictures. Go to this location, take a picture. The natural language interface, clarissa, could repeat what steps there were in the procedure for them. So they can be quite limited.
50:52
I think that you know when you, when you do code these it's not unlimited in what they're allowed to do Now as that grows more and more capable and you do code these, it's not unlimited in what they're allowed to do Now, as that grows more and more capable and you do start giving them more and more, like controlling the spacecraft, like what was done with a remote agent, then you know well, you have to think more carefully In that case. You know it's the. You know we do NASA kind of invented. You know it's the. You know we do NASA kind of invented. You know this really robust systems engineering process, which you guys have probably heard about or seen, and risk management is one of the key elements of it.
51:37
And NASA, because these missions it's not like consumer products in a way where you get a chance to test them, try them out, do limited pilots. You send one of these satellites up, you get one shot. It needs to work. And so NASA, I think, think, has done a great job with their, their, their um test and evaluation, independent testing, evaluation, uh, to to help avoid, you know, identify risks and then avoid them oh, it looks like we may be revisiting isaimov's Three.
52:18
Laws of Robotics. At some point.
52:20 - Rod Pyle (Host)
Here you go. Let's jump to one more break, and then we'll be back with Tart.
52:27 - Tariq Malik (Host)
You know, so I'm really fascinated. You outlined like several or a series of different use cases and what that? Always when I think about what you know, the different types of systems I have in the house to do a lot of different things I always think about compatibility, because it sounds like you might have an agent to check this one thing, or an analysis program that uses AI to weigh a different thing, and in space or on a spacecraft, I would assume you might instead of having one AI system, like data you know doing all of the things it sounds like the industry is leading to having really specialized either agents or AI driven systems to do tasks that then have to work together. And I'm curious how you approach at USRA, you know, looking at the compatibility of those systems, that they can understand each other and then, I guess, work together to output something that the human user is hoping to get from something like that.
53:31 - Dr. David Bell (Guest)
Yeah, that's a great question. You know one of the that the human user is hoping to get from something like that? Yeah, that's a great question. One of the research projects has been about, like the mobile agents, having these things work together with satellites. Nasa has projects where you have swarms of satellites that work with each other to achieve objectives.
54:00
I think historically with our AI models, we've always done an AI solution for a task and in many cases that are so specialized for space missions, a lot of that will continue. One of the things with the large language models and what you're seeing in evolution in AI now is where a single AI algorithm will support multiple downstream use cases. We go to Gemini or ChatGPT and we ask it any question we might ask right, we don't go to oh, here's the chat for this topic, or here's the chat for math, here's the chat for chemistry. No, we go to a single thing. There's a consolidation going on where these AI solutions are getting very broad sets of knowledge or they're integrated, where they have a single solution to support multiple experts. And one of the it's not for space exploration, but one of the things we do with space is we have satellites that observe Earth and also observe the sun, and so there's terrestrial weather, there's space weather, there's the oceans, there's the landmass for fires, there's all sorts of disasters that these satellites do. And one thing that started about two years ago and that's why, primary reason why we created the generative AI lab for science and engineering is just like this exponential growth in the size of these large language models that, in my mind, started in 2017. Just two years ago, we started to see a growth in these multi purpose large foundation models.
55:51
Using satellite data, global sets of satellite data, you can build a foundation model. Instead of having one model for wildfires or floods, or looking at rivers, or looking at tropical cyclones or atmospheric rivers or extreme rain, you create a single foundation model for, say, the atmosphere, and then you can support dozens of downstream use cases of that single foundation model. And so we're part of a NASA team that's one of the leaders in the world in doing this on Earth observation, starting to look at space weather, which affects radio frequency, blackouts, the energy grid, all sorts of things on Earth. And we've also started a team and leading a team looking at a different set of satellite data where we're focusing on weather and the data about the atmosphere and some of these models. There's a few teams. Google's got a great team that's a leader in the world. Some of these teams are getting great performance, like when we think about weather forecasting.
57:05
The traditional approach is numerical weather simulation using supercomputers where you get some observations of current state, like you have weather balloons that collect data in the vertical column, you have the satellite data and then you forecast hours into the future with simulations running on supercomputers, but those take hours to from a moment of observation. It takes you hours to come up with that forecast. Now the machine learning models now are starting to outperform those numerical simulations in multiple aspects. One, just in terms of accuracy they're getting better, more accurate in forecasting the future than the numerical weather simulations and two, in minutes from from taking observations in minutes you can create a multi hour forecast with machine learning inference compared to those numerical numerical weather simulations which took hours to forecast in the future. And so that is. You know I think it's coming very quickly that that will become the primary mode uh for weather and also for um for up for other uh uses of satellite data.
58:29 - Rod Pyle (Host)
yeah again, if, if someone at NASA or a university was foolish enough to hire me to be a data data analyst which would be a very bad idea and if I was working on, you know, one of those people that that helped to sift through reams of data, looking for exoplanets, for instance, stuff coming back from the web or from tests um, I'd be kind of worried right now if I didn't really understand what was going on, or maybe even if I did understand what's going on. But, if I understand correctly, part of what USRA and your organization does is also workforce development.
59:02 - Dr. David Bell (Guest)
Oh great question.
59:03 - Rod Pyle (Host)
You take all that into account, right.
59:05 - Dr. David Bell (Guest)
Yes, absolutely so. Yeah, our mission is research, development and education, Absolutely so, yeah, our mission is research, development and education. And so for we do a lot of student programs. Historically, one of the big projects that NASA, that USRA led, funded by NASA years ago, was for to improve engineering, and so NASA wasn't happy with the quality of engineers coming out of universities. They thought they were too theoretical, and so we ran this advanced design program where we did projects involving students and sometimes faculty, and so project-based learning wasn't a thing in engineering. There were no capstone courses, there were no project-based courses at the time, and so thousands of students and projects were run through this program. It's credited with changing the accreditation standards for engineering schools to require project-based learning, capstone projects, and so now we have two, we do both. We have a Feynman Quantum Academy, which we can talk about the quantum part, where we bring in undergraduate and graduate students and we engage them in projects and so they participate in our research teams and they help us with the research. And then we also they're learning, and we do this on the artificial intelligence side as well, and those end up becoming that's a great workforce pipeline for us, for NASA, for other contractors in the government. So this is a great way to do project-based learning and hands-on in the projects and the data sets and the algorithms that we're really interested in.
01:00:57
Another thing we started to do about five years ago was we started creating curriculum that uses NASA open data, NASA open Source and working with universities. Well, first we developed it, we taught it on the NASA campus and some of our other federal sponsors where to help upskill the workforce, and so we you know intro to we create an aviation data science course, for example, and we did an intro course, an advanced one, advanced two. But then we work with universities to do special topic courses and then so, for example, with UC Berkeley, one of our staff scientists created an aviation data science course and then taught it up at UC Berkeley and now it's an undergraduate and graduate level course that the faculty up there teach and it's sustained with tuition and everything else. So the beauty of that is that by the time we want to bring those students, they gradually want to hire them into our workforce. They already know the data that we care about, they already know the kinds of end use cases we care about and they know the algorithms you know state of the art algorithms for working with that data for those end use cases I'll give.
01:02:15
On the quantum side of it, I'll give one nice story. Well, if you're interested in quantum, I'll give you some more Well.
01:02:25 - Rod Pyle (Host)
I guess, yeah, I guess my question about quantum is really, in short form, how it will change the way we see AI today, or will it just kind of be more of the same?
01:02:39 - Dr. David Bell (Guest)
So the big focus for us, so one of there's a lot of quantum can be used for optimization problems. That's a. That's a type of problem that is, you know, is all over the place in space. We have to optimize things and we need solutions that inherently in them there's an optimization problem. And so when we started at the beginning of this, you talked about this quantum artificial intelligence lab where we did a three-way agreement among NASA's Ames Research Center, google and USRA with the Research Institute for Active Computer Science. We founded this quantum artificial intelligence lab, and so we started by USRA, installed and operated what was the only commercial quantum computer at the time in NASA's the NASA Advanced Supercomputer Center, which is at NASA Ames, and that was a quantum annealer which solved optimization problems. And so we looked for all the different types of optimization problems in space and aeronautics and we started experimenting with that and and worked to continue to evolve how optimization problems could now be used with quantum annealers but also gate model quantum computers, which is the primary type of architecture for quantum computers now, and now there are multiple different commercial gate model quantum computers and over the course since we started that in 2012, you've seen several milestones which I think will help answer your question.
01:04:24
So one of the first things was quantum speedup. Will the quantum computer be able to solve problems faster than classical computers? So is it going to be really fast at solving problems? And so Google spun up a million CPUs in their cloud. Wow, they competed against the one quantum computer we were operating at NASA's Ames Research Center, and the one quantum computer was 100 times faster than the million in solving this particular problem. So you could argue it's 100 million times faster if you were comparing one compute unit to one compute unit. And so that's just quantum speed up, and it was a real math problem. So that's just quantum speedup and it was a real math problem.
01:05:17
The next big milestone so after that, google built their own quantum computer and with them we competed against the best supercomputers at Department of Energy and NASA's Ames Research Center on another math problem to solve, and that got an article in Nature of quantum supremacy, and what that was is a problem that you could not solve without decades of computing on the best supercomputers, but you could solve it on the, the quantum computer. So that, so for one thing, you could have it on the quantum computer, so for one thing, you could have quantum speedup so you can solve things much faster. Quantum supremacy where, hey, you're going to solve things that you could never solve on classical computing. And then what the industry is working towards is quantum advantage. These are still real mathematical problems, but they're not real world end use cases of quantum computing.
01:06:17
Quantum advantage is where the quantum computers are actually solving problems faster than for a real world problem, than you could solve with classical computers, and that's what we in the industry are still working towards, just on the workforce development. Just one fun story on that quantum supremacy experiment is we have a Feynman quantum academy where we bring in students and we bring engagement projects and leading up to that quantum supremacy experiment, one of our students ended up being the lead author in the paper that described the software and the benchmark for the quantum supremacy experiment. So, and then you know a lot of those folks get hired into industry as well that we do in our work for small workforce.
01:07:07 - Tariq Malik (Host)
I'll bet they do yeah, well, you know we we've talked a lot, uh, about like applications right now, like where we are in the technology today, and obviously I watch a lot of sci-fi a lot of it bad, so I'm good, but but I'm I'm very curious, you know, as we kind of close out the the hour, if there is an AI in sci-fi that really stands out as either nailing it or that you would love to see be real. I mean, and recently we've seen great ones like TARS from Interstellar and Gertie I mentioned from Moon going all the way back to Love Him or Leave Him Hell, or Data from Star Trek, all across the board. I'm curious if you have a favorite at all.
01:07:53 - Dr. David Bell (Guest)
Oh, that's a great question. I mean Interstellar is, he had a fun attitude. So you know, one that actually has character I think is pretty cool. Yeah, so I would go with the one you mentioned. Yeah, I mean, I think mentioned, um, yeah, I mean I think, uh, you know, as you know, astronauts in space will be very isolated. You know, we've all seen that through the pandemic people working remote, you know, even kids, you know, not going to be able to go into high school, um, you know, working remote being isolated, and so we astronauts in space are very isolated. And so, you know, having an, having an AI that actually has some personality, that actually, you know, helps with the isolationism, and I think you know the human, I think that would be, I think I think that would be the kind of one I would be most interested in for the humanity of it, to help the humans as they're in space exploration, with even the non-technical part of it, which is, I think, pretty challenging.
01:09:06 - Rod Pyle (Host)
Well as the boomer in the room, I have to say my vote would be for Robbie, the Robot from Forbidden Planet, which probably a third of our listeners haven't seen, and he was an alien-created robot, so maybe he doesn't count, but he was a nice guy and we liked that. Unfortunately, our time has come to an end here. We've really enjoyed it. This has been Episode 159, that we like to call AI in space. I want to let people know you can keep up with the latest on the R I A C S at R I A C S, dot U S R a dot E D U, where there's lots of cool things going on. Before we close, david, is there anything coming up in the near or mid future that has you particularly excited You'd like to tell us about?
01:09:51 - Dr. David Bell (Guest)
well, well, we're always looking for partners. You know we work with academia, industry, government and, uh, you know there is a there is a consolidation going on in ai right now with these massive investments, so we definitely are looking for partners to help leverage all of this work we've done with the federal government and get it out into the world. We've done many projects with Google. We worked with Nissan North America to get Rover technology into managing fleets of autonomous ground vehicles, cars, and so we're always looking for partnerships. We will have a release soon on this Gen AI lab, which I'm excited about, and we'll have a steady roadmap of incremental releases after that once the first one gets off the door. But lots of fun stuff going on. It's an exciting time for AI, for sure.
01:10:44 - Rod Pyle (Host)
It sure is, tarek. You and I have to come up with a proposal to send them, so they'll have something to laugh at on a Monday morning.
01:10:51 - Tariq Malik (Host)
That's right.
01:10:51 - Rod Pyle (Host)
Speaking of Tarek, where can we find you indulging in neural networks these days?
01:10:56 - Tariq Malik (Host)
Well, you can find me at spacecom, as always. You can find me on the Twitter, I guess. Well, x, right and Blue Sky at Tarek J Malek and, of course, tonight you'll find me on YouTube at SpaceTronPlays, because it is a new season of Fortnite and it is all Star Wars all the time. This weekend is free comic book day and Star Wars day. Lots of great stuff all happening. That's not about space, but it is exciting.
01:11:22 - Dr. David Bell (Guest)
Well, thank you so much for the time. Appreciate it Exciting to you.
01:11:25 - Rod Pyle (Host)
Tarek Likewise. And, of course, you can find me at pilebookscom or at astromagazinecom, as always. And, uh, remember, you can always drop us a line at twist at twittv. We welcome your suggestions, ideas, comments and your adoration um new episodes of this podcast published every Friday on your favorite pod catcher. So make sure to subscribe, tell your friends, give us reviews good ones, please, cause we're counting on you. And don't forget, we're also counting on you to look into joining Club Twit this year. Besides supporting Twit, you'll help keep us on the air and bring you great guests and horrid space jokes, because that's what we're all about here. We're also now featuring annual subscriptions again, so if you're interested in that, that is available to you. And David, thank you very much for joining us today. You've been patient with our questions. We always appreciate that and I hope we can have you back sometime. Thank you so much. Really enjoyed it. Thanks, take care everybody. See you next time.
01:12:24 - Leo Laporte (Announcement)
The tech world moves fast and you need to keep up for your business, for your life. The best way to do that twittv. On this Week in Tech. I bring together tech's best and brightest minds to help you understand what just happened and prepare for what's happening next. It's your first podcast of the week and the last word in tech. Cybersecurity experts know they can't miss a minute of security now every week with Steve Gibson. What you don't know could really hurt your business, but there's nothing Steve Gibson doesn't know. Tune in Security Now every Wednesday.
01:12:59
Every Thursday, industry expert Micah Sargent brings you interviews with tech journalists who make or break the top stories of the week on Tech News Weekly. And if you use Apple products, you won't want to miss the premier Apple podcast Now in its 20th year, macbreak Weekly. Then there's Paul Theriot and Richard Campbell. They are the best connected journalists covering Microsoft and every week they bring you their insight and wit on Windows Weekly. Build your tech intelligence week after week with the best in the business. Your seat at tech's most entertaining and informative table is waiting at twittv. Subscribe now.
Share: Copied! This Week in Space #159
May 2 2025 - AI in Space!
USRA’s Dr. Bell on Robots, Rovers,… All Transcripts posts Contact Advertise CC License Privacy Policy Ad Choices TOS Store Twitter Facebook Instgram YouTube Yes, like every site on the Internet, this site uses cookies. So now you know. Learn more Hide Home Schedule Subscribe Club TWiT About Club TWiT FAQ Access Account Members-Only Podcasts Update Payment Method Connect to Discord TWiT Blog Recent Posts Advertise Sponsors Store People About What is TWiT.tv Tickets Developer Program and API Tip jar Partners Contact Us