Can AI Read Your Mind? Neurotech and Privacy

Transcript generated automatically from the episode audio and lightly cleaned up (filler words removed, misheard names corrected where clear). Speaker labels were added afterward and may contain errors.

Rob May: Hi, I'm Rob May. I'm the co-founder and CEO at Neurometric.

Anna Kirk: Hi, I'm Anna Kirk. I'm sales lead at Thread AI.

Rob May: And this is the AI in NYC show. So, welcome everybody. Ryan is out today, so you just have Anna and me. And our guest today is Kristen Mathews from Cooley, who is a privacy lawyer and has particular expertise in neurochnology. So, we're going to talk a lot about neurochnology, brain machine interfaces, things like that, and how they apply to, the AI industry and how that's going. So, Kristen, welcome to the podcast.

Kristen Mathews: Thank you. Happy to be here.

Rob May: Yeah. So, let's let's talk a little bit about a lot of people probably don't know that even privacy law is like a specialty and what that would mean like. What does a privacy lawyer do? What what do people come to you with? What kinds of issues and problems? And what's your main job?

Kristen Mathews: My main job is to assist businesses of all shapes and sizes to comply with the maze of privacy and cyber security laws that apply to their business. So let's imagine, I don't know, a B2C business of some kind that naturally in the course of their business has information about consumers and they have to collect it, use it, and share it in a way that complies with laws. And that's what I help them to do. And I've been doing that for my whole legal career, which is almost 30 years. But recently, I found this niche, neuroch, and realized how important it is from a privacy perspective given what neuroch is capable of doing and decided to zero in on that and created this little subsp specialcialization for myself, which I'm greatly enjoying.

Anna Kirk: How would you define neuroch?

Kristen Mathews: Neuroch is technology that detects electrical and other kinds of activity from our brain. And from this kind of information, it can accomplish things. Maybe it can accomplish something like decoding feelings, for example. If it gets to know your neural data, maybe over time it can learn to decode it into what you're feeling at the time. Or maybe a sense of rec recognition or maybe even some speech, some inner speech or inner thoughts that you're having. So neuroch essentially is tech that can detect brain activity and learn from it.

Rob May: Yeah. I there was a whole big wave in like 2018 through maybe 2021 when deep learning came about where all these people so one of the problems with neurochnology and measuring that is like your skull actually doesn't pass signals very well through it right so it's hard to read this is why like a lot of your brain m like your best ones are very invasive they actually put them down in your head and I know all this because I sat on the board of a company called neurable for a while was doing some of this but the so the whole idea everybody had was like oh we'll just use machine learning to decode these brain signals and figure out what they are and what they're saying. It turns out it's actually harder than people thought. But with like Neuralink and some of the other things that have have have really started to push I feel like they have pushed the industry forward over the last couple years from a technology perspective and it's becoming better and better and better. Actually, I'm I'm an investor in a company out of Boston that slaps a it's called Beacon Biosignals and they slap a little thing on your head and they can just they can measure enough of your bra. So, EEG is a good EEG is your brain, right? Not EKG. Okay. I was think EKG is heart.

Yeah. So, EEG is many many diseases have a EEG manifestation, but doctors don't like to hook you up to it because it's a big pain in the ass and it's expensive and it takes a while. And so this thing like triages it and it's basically like is there enough there that we should hook you up to an EEG, which is kind of cool.

Kristen Mathews: So you're right. There's a big difference between what we're calling invasive neuroch versus what I'll just call wearable neuroch because the skull it happens is a barrier to that electrical and other activity in the brain. And so the fidelity of the information that you can glean with the skull in the way is less than if you have it invasive, which is which means that's it's actually implanted with neurosurgery into the brain. So there's there's a couple of different kinds of invasive neuroch also. So it could be that you full-on neurosurgery implant an electrode into the brain. But there are less invasive but still invasive methods of getting this neuroch into your brain. For example, one company has one that is inserted through a blood vessel and another company has one that is like an electrode panel that is a layer that you place on the surface of the brain but under the skull.

Okay. So these are three invasive neuroexs because they're under the skull but they are three different levels of invasiveness and again once you get beyond the skull the fidelity of the data improves and therefore you can do more with it and then there's the wearables like you mentioned the company in Boston those come in the form of headbands helmets ear attachments even bracelets because remember we have the central nervous system and we have the peripheral nervous system. And so here right on your wrist is your peripheral nervous system. And so there are bracelets that are tapping into some information that they get can get through your muscles.

Rob May: Do we So you know part of the focus of this podcast is New York City. What is the where is the New York or I'm sorry where is the neurochene in general? Like is New York a hub for this or

Kristen Mathews: New York is very much a hub for this.

Anna Kirk: When I think about LLMs, and granted I'm I'm no neuroch expert, but when I think about LLMs, I think about kind of models trained on text or images or maybe even videos, but I'd love to hear kind of a little bit more about your perspective on like what you think this kind of new AI wave has unlocked in terms of what can we detect and maybe what are what are different patterns that we can map from brain activity and how is that actually translating to real innovation today?

Kristen Mathews: I'm going to take you back 100 years to answer your question.

Anna Kirk: Okay, great.

Kristen Mathews: The in 1924 a scientist realized that our brains have an electrical element to them and that was the first EEG detection. And so that means that a hundred years have gone by during which this tech existed, but the data that we were getting from this kind of tech wasn't really able to be decoded very well. And it's only within the last few years because of AI that we've been able to take all of this neural data we're calling it, apply AI to it, and learn a lot more about a person from it. And so it's all about AI really that has advanced this technology over the last few years. AI is involved in all of this in a few different ways. On one hand, AI is used to decode neural data that is gathered by these various technologies. EEG among them, but frankly, EEG is just one of many different texts that are able to detect brain activity. But in all cases, AI applied to this very vast array of neural data can be used to do things.

Maybe it can be used to discern whether a person is going to have a seizure within the next 20 minutes. And by the way, I didn't mention this earlier, but neuroch not only detects brain activity, but some of it also stimulates the brain with an electrical pulse to have an effect on the brain. So with this technology that I mentioned seconds ago that can read neural data and learn something from it. For example, maybe this person is about to have a seizure, they could also stimulate the brain in response to that knowledge to suppress the seizure.

Rob May: Yeah. I like so note to our producer Madi, one of my favorite articles of all time is this article from probably the mid 2000s, like 2005, six, seven, eight, somewhere in there. Called the god helmet or I'm sorry it was about this thing called the god helmet and it was called I still remember it because it had such an impact on me it was called savant for a day and they took what's called a transcranial magneto stimulator right so you mentioned earlier brains produce magnetic waves right because when you have electrical waves the way those things propagate through empty space is electrical waves create magnetic waves and vice versa and everything else and so one can create the other so what they do what they did in here was they had this journalist like draw a picture of a dog and it was like a crappy dog like like I would draw because I zero artistic skill.

And then for 20 minutes or 30 minutes, they like zapped his brain with these magnetic waves to stimulate the electrical activity in the art drawing part of your brain. And he drew an amazing dog. And then they had him draw it like every couple of minutes as the stimulation faded away, the draw dog got worse and worse and worse. And so it's really incredible. And I think there's a there's a tool on the market now, right, which will stimulate like for athletes and stuff that's similar to maybe maybe TMS you have to have a lab, but you can't just buy it. But there the related technologies.

Kristen Mathews: So there's a lot of direct to consumer neural tech as I mentioned earlier. Some of it is invasive and you need to go through a doctor or participate in a clinical trial in order to get it. But some of it is wearable and some of that wearable tech is direct to consumer. You can buy it on a website and several of them are being used to help people sleep better or to meditate better or possibly to you know impa enhance the functioning of their brains.

Anna Kirk: What are what are the ethical implications of that when you think of like and again coming at this from a complete novice but what are the ethical implications of that if you now have the kind of technology that can not only pick up on certain patterns but then also stimulate different regions of the brain if that is at all like AI-driven or autonomous like who is culpable for something going wrong.

Kristen Mathews: It's a really good question and that's probably why I'm at the table. Most and I'll even say all of the neuroch that I'm aware of that's actually out there is all good and I don't see any business using it for anything that any one of us would be concerned about. Using it to again help somebody who has ALS and therefore can't communicate on their own to communicate with their friends and family. Helping somebody who's paralyzed to be able to move their arms with an exoskeleton. Helping somebody who has epilepsy or who has had a stroke or who can't see very well to see better. All of these things have no ethical issues because they are exactly what we should be doing with this technology.

I think the ethical issues that people are afraid about are first of all theoretical. And maybe we can just talk about a few. Let's talk about inner speech. So ALS people who have ALS and can't communicate with their families, this technology is enabling them to communicate with their families. And what the tech is doing is it's decoding their inner speech. So let's think about that from a pract.

Yeah. But let's just be really careful with what we're saying because there's more than one kind of inner speech. There's intended speech. That's the term of art that we're using. That's when a person is trying to speak but can't. And so some of this neuroch is tapping into intended speech, decoding it, and enabling that speech to show up on a computer monitor, which is a wonderful thing. And there's no privacy issues with it, right? Because it's intended speech. The person wanted to speak, they couldn't. This tech helped them do what they wanted to do. So there's no privacy issues there, right? So let's talk about a different kind of inner speech, which I'll just call inner speech. That's more like this the speech that's happening in your mind that you're not trying to say. That's where the possible privacy issues come in because this tech is gathering and decoding speech that you didn't try to say. And I think that's where we might have some, you know, not ethical issues, but we want to pause and figure out how are we going to deploy this tech in a way that is safe and fair and best for humans and best for humanity.

Anna Kirk: Is there a very clear and defined line between intentional speak and maybe internal monologue? Like how is that when you think about the e ethical considerations of decoding intentional speech? Like in theory that sounds that sounds very clear and defined, but are are there gray areas and kind of decoding which is which?

Kristen Mathews: It's a really good question. I don't know the answer to that question. I don't know if it would be easy and very clear bright line for the re the tech to kind of only grab Yeah. The intended speech and not grab the inner speech. I don't I don't know the answer to this question. It's more for these researchers, I think, to answer.

Rob May: Yeah. I mean, I would be worried about that because I think one of the trends we've seen in tech, we've talked about this on the show before, is like constantly lowering the friction to communicate. And in the tech world, people believe lower friction is a good thing. Always. Friction is nature's way of making you pause and think. And so, to give you an example, 25 years ago, if you wanted to express your opinion, it'll write a letter to the editor and hope it got published. And then web came around and you could maybe do some stuff, but it was hard to publish on your own. But if you wanted to learn HTML, you could do it. And then tools made it easy to publish. And then Twitter came along and it was like just whatever thought you have in the moment, just put it out there. And it's like the as the barrier's gone down, I feel like the dialogue overall has gotten worse.

The sort of public discourse around these things because I mean how how many times like I don't know if you guys tweet like I did a lot in the early days and there were times where I started to like would type one and then I'd be like not going to send that, right? Because a half hour later you don't feel so angry about whatever somebody just tweeted. And so like friction can be a good thing. And I like I worry about a I worry about a world where companies buy that data and just advertise to you based on what your selft talk is or whatever your your inner voice is and you know you're you have a you don't want to play the lottery or you don't want to go to the casino or whatever whatever your vice is, right? It's like you're thinking about it and you're debating it and they're showing you an ad for it because that's

Kristen Mathews: I worry about that world too and that's probably number two on our list of ethical issues. I want to add though that the researchers who are looking to decode inner speech as opposed to intended speech are also looking for ways to put some controls around it. And some of them are experimenting with a I'll call it a password, but it's basically a string of words that you would think in order to signal to the tech not to proceed with it with its decoding.

Anna Kirk: Interesting. Then I Okay, so I always have the problem where if I'm ever like don't think about tacos, the only thing I can think about are tacos. And I wonder if there would be a similar I'm sure there are many ways around this.

Kristen Mathews: It's more like think this string of words and the neuroche will stop its decoding.

Anna Kirk: Interesting. Okay. So it gives you kind of like the ability to kind of like pause command essentially. Yeah. Okay. That's fascinating because the researchers realized exactly the concern and the concerns that we're talking about right now.

Kristen Mathews: Yeah. Now you mentioned what I'll call ad tech. I mean, as I said earlier, I've been in the privacy field for a long time, and one of the very early things that I dealt with was behavioral advertising online, and it was creepy to people back then. Now, it's totally normal. But let's think about advertising to someone based on their neural data because they happen to be wearing a headset at that moment. And let's take it a step further. Let's think about political advertising at a time like that. What kind what what should the rules be for this sort of tech?

Rob May: Yeah. Particularly because something like political advertising is designed to enrage you, right? And I think we could argue people don't need to be any more enraged than they are.

Kristen Mathews: So speaking of enraged, in 19 in the 1960s there was a researcher and his name was José Delgado and he was associated with Yale as I recall but he I guess had an affiliation with Spain and he was a neuroch researcher and he went to Spain and he found a bull and when I say bull I mean like the animal and with others they performed neurosurgery on this bull. And they implanted an electrode into the bull's brain at a particular location that they had figured out in advance and they stitched him up and brought him back to consciousness, put him in an arena and José held a red cape and he had a remote control in his hand that was connected to this inter this implanted electrode in the bull's brain. Red cape. Bull comes charging, pushes the button on the remote remote control. The bull went do like not docile, not in pain. He just stopped charging and looked around and then he releases the button. The bull comes back. So, can you imagine like what that means for the capability of this sort of thing?

Rob May: Yeah. Or have you ever heard this study? Do you know the term wireheading?

Kristen Mathews: Mhm.

Rob May: Yeah. So there was a there there was a study where they took these rats and they implanted electrodes in the pleasure center of their brains and they gave them the ability to sort like hit this lever and stimulate that and it gave you a rush of like dopamine or whatever and all the rats did was hit the lever until they they like didn't eat, they didn't sleep, they just pounded the lever and it's another good example. Yeah. And that those are things that you'd have to be careful of. Let Oh, well I was Do you have another question?

Anna Kirk: No, I was so a slight tangent, but I'm curious so when we think about like how legal precedent in the space has evolved over time as kind of our ethical considerations around privacy and data have shifted. Do you think that the same kind of best practices still stand like can we pull on precedent that was written 20 years ago or is the complete like paradigm shifting?

Kristen Mathews: It's a very astute question that you're asking. And my answer is that I don't think that privacy laws that are already on the books work very well for this technology. We have decades of privacy laws that are intended to regulate data like names, email addresses, social security numbers, health information, I don't know that sort of information that I would call data. And it exists outside of you. It exists on paper and it exists on computers.

It's different from our thoughts and our feelings and our emotions. I mean, I didn't mention earlier, but some of this tech can decode emotions or likes and dislikes of products. And the privacy laws that are out there, which are more geared towards data, aren't really designed for this. And I don't think they're working that well for this. And in fact, the course that we're on right now is to take those laws that have been on the books for years and to throw neural data into those laws by literally creating a definition called neural data and saying, "See this law and see all of its requirements. Well, comply with that for neural data, too." That's what we're doing right now in the law. We're just adding neural data to these old laws. I don't think it's working that well. And I'll tell you why I don't think it's working that well. This technology is good. I told you everything it's doing for people. And if we were to just say, hey, for all neural data, you have to do all of these things. You businesses who are in the neuroex have to do all of these things. Most of them are completely unnecessary for most of the tech that we've described. And it's burdening that tech. It's burdening the companies and therefore it's taking the tech and the solutions away from the humans and it's unnecessary because there's nothing wrong with this step.

So I think that it would be better instead of saying hey you have to comply with all of these requirements for all neural data. I think it would be better for us to come to a table like this one and think about what are we really concerned about. When I told you about this tech I you yourself realized that there are some uses of this tech that we might want to regulate or put some rules around. Let's all identify those and let's isolate those and let's deal with those without encumbering the entire industry with laws that aren't necessary. That's my opinion.

Rob May: So what's interesting is there there's talk now about taking a lot of this neural data and doing what we did with AI data and the way that we've made large language models which you know another word for those is like foundation models. So foundation models is a category. Large language models are a type of foundation model which means they have lots of data in them. They can do lots of things and other things are built on top of them, right? People are talking about building brain foundation models with some of this data. And so I guess what that would do is the way you would give an image to an LLM and ask it what's in the image and it could tell you could either maybe give some words or an image to brain foundation model and it could tell you the brain patterns that would be activated or vice versa and that could really be and then that could give this power to lots of people to understand what's happening and also could be great because maybe you could find things about disease and figure them out and have medical use cases and all that, but what's the where does that stand in terms of like both privacy law and whether or not people are actually pursuing it?

Kristen Mathews: So, it's being pursued. And as you said, whereas an LLM is trained on words and images, a brain foundation model is trained on neural data. And when I say neural data, I'm talking about the EEG data, the fMRI data, all of the kind of data that's been collected by hospital systems over the course of decades as well as in the context of clinical trials as well as these wearable private wearable B2C products. This neural data can be used to drain and create a brain foundation model just like an LLM, but not an LLM. It's brain. I don't know how to describe it other than a simulated I don't want to say simulated brain because it's probably an overstatement but an AI an AI model that mimics a brain and can therefore predict how it will react to different things and I think disease progression is a great example of how a brain foundation model would be used.

Anna Kirk: It's interesting you can you can talk about like like outputs of that that address conditions or maybe other neurological disorders and then I can also think of elements of that which you mentioned earlier where it's like okay you can feed in a stimulus and it might it might be able to ingest different brain activity around like oh like this is the emotion that you're feeling and that varies so much from person to person. I'm curious kind of and this is maybe a broader kind of open-ended question for the people who are watching who are more researchoriented. Maybe we have Kate on the show in the future and she can tell us the answer to this as well.

Rob May: Well, Margot's coming on. So Margot Hanley is one of our friends who's Yeah. Very deep in this space. So she's going to come on and talk about this.

Kristen Mathews: She should. In fact, she is the queen of brain foundation models and she's writing an amazing manuscript on it right now. Definitely speak with her about it. Yeah. But you mentioned different for different people. So let's talk about that because it's connected to the legal issue here. Researchers are experimenting with whether neural data of all different kinds can be used to identify a person just like a fingerprint. Okay. So it turns out that our the patterns of our brain and the way that different portions of our brain communicate with each other is unique to each of us. And if you look at someone's brain enough, you can identify them using it. Now this doesn't mean that you can identify someone if you have no experience with them. I mean you have to be able to compare the pattern you're seeing today with the pattern that you already collected about the person yesterday.

Just like a fingerprint would be. If you got a fingerprint and they weren't in the database, you wouldn't be able to identify them. But if you did have their brain pattern from the past, you could identify them. And researchers are finding ways to do this for and as an example, if it works, you'd be able to use your neural data as a password to get into a computer. So what that means though is that neural data is inherently identifiable. It's a biometric. Okay. Which means that it's covered by a host of additional laws that are specific to biometric privacy just like fingerprints and retina scans and face facial recognition.

Anna Kirk: And so do you think this will be the case? My my perception looking at both web and then AI data related privacy laws and forecasting that in neuroche is that Europe has actually been a leader here. They've been more proactive maybe than the United States whereas the US has been more responsive. So, so I guess two two part question for you. Number one, is that actually the case or is that a misperception? And then secondly, are there differences in what the two areas of the world are doing for neuroch?

Kristen Mathews: So you're correct, Europe has been more proactive in regulating AI. As far as I know though, they haven't enacted anything specific for neural tech. I believe that what they're doing is they're saying, you know, our current laws, the GDPR, which is about privacy, and the AI act, which is about AI, they apply to neuroche too. We don't need to write something special for that. Naturally, they just apply. So, that's what I believe Europe is doing with this. The US at the federal level isn't doing anything. But at the state level, states are actually enacting specific laws that have the word neural data in them as opposed to just saying our current laws apply. Four states have done this so far. Montana, Connecticut, Colorado, and California. Did I say California already?

Anna Kirk: I don't think so.

Kristen Mathews: Okay. So, those are the four. And the way that they've accomplished it has been to take a pre-existing law that was a privacy law and it covered any kind of personal data and they've said this law now has this new definition of neural data and we are going to consider neural data as defined to be what they call sensitive personal information which gives it enhanced regulation. So if data is not just a name but it's but it's sensitive a social security number, racial information, health information, neural data, biometric data, that makes it quote sensitive under the law and therefore businesses that process it have to do more. They have to get consent in order to process it. They have to do data protection impact assessments which are basically selfassessments that a business has to do where they evaluate how they're using the data, where they're getting it from, who they're sharing it with, how long they're keeping it, what are the risks involved in it, how am I mitigating those risks, and taken all of that into consideration. Should I be doing this? That's a risk assessment.

Anna Kirk: You mentioned that four states have started to adopt this. They have adopted they have adopted this kind of labeling of neuro data as as sensitive. Are do you see most states kind of adopting a similar framework or do you think it'll vary state by state and then again what are the jurisdictional impacts of that?

Kristen Mathews: So, states tend to copy each other, and we're already seeing it happen because three of them have like two of them have copied one other. Montana, by the way, they did it differently. I can explain why if you're interested. But I and there's nine more. As of the last I checked, there were nine bills that would regulate neural tech specifically that were introduced in 2026. So far, there are also a bunch of bills that were are still on the books from 2025 where their legislators are still in session and therefore those bills are still alive technically.

So, there's quite a few bills pending that would specifically regulate neuroch and they take all different methods like there's one that is specific for employers. So, think about what an employer can do with neural tech, right? That's one of the things we were talking earlier about. What are the ethical issues here? And we never really got through our list, but another one that would be on the list is employer use of neuroche to monitor employee attentiveness, productivity. So legislators feel that this is something that is worthy of regulation. So some of these bills I mentioned to you are specific to the employment sector. Some of them treat neural data just like they treat genetic information. Illinois is a good example of this. Their bill will take their genetic information privacy law and simply add neural information to it. And by the way, that's what Montana did, too. I told you earlier Montana was different than the other three. It's because they chose to put neural data not into their privacy law, general privacy law. They chose to put it into their genetic information law.

Anna Kirk: And what is the implication of that in practice or how have you we seen that kind of play out?

Kristen Mathews: The implication of the choice as to which law you put it in?

Anna Kirk: Yes.

Kristen Mathews: Yeah. Well, I mean the privacy laws are more general and they apply to privacy to personal information generally and they require all kinds of general things. So, I guess throwing neural data into those is going to cause businesses to have to comply with a whole bunch of different general things. I can give you some examples like you have to have a privacy policy that articulates what you're collecting, what you're using it for, how long you're keeping it, who you're sharing it with, all of that. You also have to get consent for processing neural data. You have to do these data protection impact assessments that I mentioned. You have to give give people the right to access their data. So, they have to be able to call you up or email you or fill out a form and say, "Hey, you have my neural data. Give me a copy of it." Or, "You have my neural data. It's wrong. You have to correct it."

Anna Kirk: Hm. Okay. That's a

Kristen Mathews: It's called the correction, right? And it exists under these general laws. You have to delete someone's neural data if they ask you to. So, these are all kind of general privacy laws. And so, any state that just throws neural data into those laws, it's going to be all of that stuff. If you choose to throw neural data into a genetic information law, it's going to be different. Genetic information laws say things like, first of all, consent. You can't gather it without consent, but they also have specific provisions about sharing it with insurance companies, employers, and there's a third category actually of where who you can't share it with. It's insurance company. I don't remember what the third character, but so it's so they don't want insurance companies to determine your rates based on your neural data or your genetic information.

I you asked me earlier whe whether I think that existing privacy laws are suitable for neural data and one of the ways where I think they're not is in this point about deidentification because I don't think that the methods that we use to deidentify typical personal information like names, email addresses, and social security numbers is the same as how we would deidentify neural data. And if so, if we were to apply those laws and try to deidentify neural data using those laws, I don't know that it would be effective.

Rob May: Cool. We always end on a New York related question and we asked you when you were coming in, you live a little bit outside the city and so we were saying like when you come into the city, is there a place you really like to go? And we decided we would talk about our favorite places to go for pasta. So, do you want to start?

Anna Kirk: I actually, you know what? I told myself that I would think about this during the episode and like can I go last?

Rob May: You can go last.

Anna Kirk: I need to go last. There are too many good ones.

Kristen Mathews: So, my go-to is L'Artusi. Yeah. And I'll share that when we go, we order all of the pastas almost regardless of how many people are at our party.

Rob May: That's the way to do it.

Kristen Mathews: So that we can just try all of them.

Anna Kirk: They do make excellent leftovers.

Kristen Mathews: Yes. I will say also. Absolutely. And for the record, we also order every dessert. Oh, in fact, my husband made my day one day when we went to a restaurant and the dessert menu came and he just ordered every single one of them.

Rob May: That's awesome.

Kristen Mathews: The way to a woman's heart, I think, is for me. Yes.

Anna Kirk: They have a good olive oil cake, don't they?

Kristen Mathews: Yes, they do. They sure do. I'm so hooked on olive oil cake.

Rob May: Mine is probably Via Carota down in West Village. Okay. It's a really good little place. Yeah.

Anna Kirk: Gosh, there's one that I cannot for the life of me remember the name of. It is on Lafayette, like kind of in no Austria something.

Rob May: Is it Austria Caruna?

Anna Kirk: No, that can't be it. Oh, this is going to bother me and now I can't think of the name. That would probably be like my top one, but I'll have to like edit back in what what the name of that one is. Okay. Just have to say it.

Rob May: Send a voice clip to Madi.

Anna Kirk: Okay, Madi, I will send a voice clip to you of whatever this one like godforsaken pasta place is that I can't think about.

Rob May: Yes. Osteria Morini.

Anna Kirk: Yes, that's the one. Wow, you were really good. Osteria Morini. Yeah.

Rob May: Well, I found that and this is a good way to wrap up the show. So Kristen, thank you for being on.

Kristen Mathews: My pleasure.

Anna Kirk: Definitely if if you if you need a privacy lawyer, Cooley is a ve very big firm.

Rob May: Yeah, we have so many venture funds on but I actually looked this up. So we have a new website out new yorkgbt.com and I actually looked up Osteria on NOHO and I got it. Wow. Yeah. Well, it's very it's sort of like ChatGPT but trained on New York neighborhoods, things that are New York related. And yeah, there's some work to do on it, but check it out if you get a chance. Please rate us, like us, check out our other episodes, and we'll see you guys next week.

Anna Kirk: Great. Thanks, Kristen.

Kristen Mathews: My pleasure.

Rob May: BePresent is an app started by two brothers who spent too much time on social media, and so they use the same techniques that social media uses to keep you hooked. They use those techniques to keep you off your phone. Check out BePresent, and you know, you'll be a better person for it.