Russ was diagnosed with bowel cancer in late 2021 and simultaneously with smoldering myeloma, aged 40. The smoldering myeloma has been inactive; the bowel cancer has progressed through multiple surgeries (bowel, liver, lung) and is now stage 4, on active chemotherapy. He runs AI for the business he works for, so his day job is adjacent to the technology. He blogs publicly about his disease at fcancerwith.ai and on LinkedIn. He is British; cared for by the NHS with some private care around the edges. He is articulate, technically fluent, and willing to pay roughly £200 a month for AI subscriptions.
[00:00:06] Dear listeners, welcome to Faces of Digital Health and a special series called The Agentic Patient, which is a series about how real patients are using AI to navigate their health, such as finding insurance, managing symptoms, tracking disease and more. In this series, we go into details. We talk about which tools patients use, which prompts, what's working, what isn't.
[00:00:32] These discussions are intended for informational purposes only and should not be relied upon as a sole source of medical information or a substitute for your professional medical advice, diagnosis or treatment. Always consult a qualified healthcare provider regarding any medical concerns or decisions.
[00:01:01] In today's episode, you're going to meet Russ Reed Barrow. Russ was diagnosed with bowel cancer at 40 and at the same appointment with smoldering myeloma. He was handed a paper journal to track his chemotherapy symptoms, but he threw it away and instead built his own system of AI tools.
[00:01:22] He's running several cloud projects, uses a notebook LM workspace and a mobile orchestration tool cloud dispatch. And this is what we're going to dive into today. What are his key warnings, key insights and key tips that you might use as well?
[00:01:45] Before we begin, if you will enjoy the show, also check out the newsletter, which you can find at FODH.substack.com. I'm going to add the link in the show notes. And also, if you like the show, do share it with your network, share it with other patients that you might find it useful. And if you have a story to share, ping me on LinkedIn and let's talk. Now let's dive in today's discussion.
[00:02:23] So Russ, hi, and thank you so much for joining me on this discussion for a special series of episodes on faces of digital health called the agentic patient, where we're exploring how different patients use AI, what are their experiences, what are some of the best practices with the aim of informing the wider public, the agentic patient, what are some of the best practices that you might find out of the world. And we're going to talk about what can be useful about AI and where we do still need to be mindful as patients.
[00:02:53] You have a very particular story because you built your own AI system for tracking your cancer treatments. But let's start at the beginning. What's your story? When did you start using AI and when did, you know, your medical situation start, if you don't mind sharing that? Yeah, of course. I was diagnosed with bowel cancer at the back end of 2021.
[00:03:19] So just after coming out of COVID here, I think, and all the issues of that around that, I was actually at the same time diagnosed with another cancer called a smoldering myeloma, which is, yeah. So two very rare for my age. I was 40 at the time. And one of them is pretty inactive. The smoldering myeloma just sits in the background and hasn't actually caused many issues.
[00:03:45] The bowel cancer has caused me loads of issues from it started. I've had sort of various rounds of different types of chemo. I've had bowel surgery, liver surgery, lung surgery, and now I am on chemo. So it has progressed. So it's currently, so it's stage four now, which isn't ideal. But to be honest, the chemo I'm on now is pretty manageable.
[00:04:10] And I don't know, I just try and stay as positive as possible, keep myself busy. And so that is that sort of in a nutshell, the cancer situation. So I'm currently on on chemotherapy at the moment. That is every couple of weeks. And I have a rough week, which is actually this week, and then a week off and then I'm back again. So I should have a I should hopefully have a few months off in a cut in about a month or so.
[00:04:39] But that's down to my oncologist. When did you start using AI to get additional information about the treatments, about the cancer? Just tell me, how did you use it? What motivated you? Yeah, I first came across AI in my job, really. I am part of my role at the moment is I look after AI for the business that I'm part of. So I got into it initially through work, largely started with ChatGPT.
[00:05:08] I think most people probably did. Realised quite early on how useful it could be. And I think, I guess, I remember the first time I had a round of chemo, I did six months of it. It was and it was awful. And I was given a book by the hospital to track my experiences in. In fact, you're given for anyone that's got cancer, you're given so many books and so many pieces of paper over the years. So I got to the point where I didn't really keep anything. I've thrown so much away.
[00:05:37] And that was one of the many things that I didn't use. I tried it and I just didn't get on with it. I didn't find it very useful. And also, I didn't find it that useful for relaying stuff back to my doctor or my GP or whatever. And I guess when you have chemo, you quickly realise that you go through two weeks of it or three weeks or whatever the cycle is. And there are lots of things that happen again and again. And you lose track of them if you don't write them down. And there's lots of patterns that you could pick up.
[00:06:07] But writing in a book for me was never going to work. And I'm pretty sure my oncologist would never look at that book anyway. My nurse team either doesn't have time to look through pages and pages of stuff. So I started using, I had the idea then of just building a like ChatGPT. All of the LLMs let you build your own personal versions of what they do. And if you use a paid model, the information that you put into them is pretty secure or it should be secure anyway.
[00:06:36] So I think that's the first thing to bear in mind for anyone using AI for this sort of thing. You do want to be using a paid model. You don't want to be using anything free because they can train their data on it and it's out of your control a little bit. Why would that be a problem? In a way, because then other patients maybe get better answers. Yeah, maybe. I guess it's because it's personal information and it's how open you want to be about it.
[00:07:02] I think it's a good point. I, I guess I wouldn't mind it if it was used in that way. If you knew it was being used for cancer related stuff, I'd happily reshare all my information. But I would be wary of it being shared for sales related advertising information, all that sort of thing is what I'd worry about. Whereas actually I've been considering because I've got my website, which is F cancer with dot AI, which I have recently updated and that's got a lot of information on it.
[00:07:31] But I've previously thought about putting all my health information on there, like really opening up. I've got loads of data about me and sticking it on there because you just wonder if someone could do more with it than you can yourself. But anyway, so yeah, I've done, I've done lots of things. So what started with me using like a custom chat GBT to just track symptoms?
[00:07:51] I use that for my whole, the last round of chemotherapy, I used a mix of chat GBT and another AI service called Manus to daily, like I used a, I tracked my symptoms twice a day on chat GBT. And then I plugged them into Manus, which is a different sort of LLM, which I used as like a backup research bank to keep all the logs in.
[00:08:14] It was really useful while I was using it and it helped me figure out when I was going to have good days, when I was going to have bad days, why things were happening, what I perhaps forgot to do from last time around, earnings, what tablets to take when. And it showed up loads of patterns. It taught me an awful lot about how to improve how I manage it. And then there were also some massive lessons learned like LLM's chat. Sorry, I keep saying LLM.
[00:08:42] That just means any, any chat bot, anything like chat GBT or Claude or Gemini, none of them are very good for long-term memory. So if you use one tool, you always need a sort of hard backup of what you're putting into it, because I basically got to the end of my cycle, sorry, to the end of the last round of chemo, when I should have had something like three months worth of data. And between the two of the, between the two chat bots, they lost about half of it.
[00:09:12] And it's quite hard to know why there's this ongoing issue with LLM's and long-term memory. So that was a major learning. And this time around, I've done it slightly differently. I've built like a, again, there's, I detailed it on my website, but I use Claude as my daily symptom tracker. I just track once a day because tracking twice a day was a bit excessive. And every day I cut and paste what Claude gives me into a Google document.
[00:09:40] And that Google document acts as like the long-term memory. And then Claude can always refer back to it. And then I can use that Google document. I put that into another service called Notebook LM, which is a, another Google tool. And if you put it in that, that can look over all my data and it can turn it into graphics and presentations and reports. So it's really handy for, if I have a call, like next week, I've got a call with my oncologist.
[00:10:10] I've not spoken to her in a month. Ask me what's going on. And I can go into that Notebook LM tool and just say, you know, give me a summary of what's happened over the last three weeks or four weeks. How are clinicians reacting to this? Like when, do you, do they know that you do all this tracking with AI? What's their thoughts, responses? Yeah, it's difficult. I think, I think they are, I think my oncologist has been very open to it.
[00:10:40] I think she was quite surprised by the amount of information it can give you. I personally hate it when you go and see anyone. The other thing is, for instance, my, my oncologist has been away on holiday, which happens. They go away for a month at a time, which is fine because I rely a lot on the nursing team anyway. But when she, but I'll have a, I've had a stand in oncologist for that period of time. You know that they're never going to be up to speed.
[00:11:06] If you have a call with them, you'll be lucky if they've got time to read the latest notes, let alone have a full picture of your background. So if you've got a way of making it easier for them, or the other thing I've had over the years is that I've flipped between specialists. I've had a lung specialist, a liver specialist, and they all need to know what's going on. And, and actually they all don't have, it's so rare that they have the clear picture because different hospitals don't have access to the same systems.
[00:11:36] It takes the NHS, luckily at the moment I have a few things done privately, but the NHS seems to take two weeks to get a scan from one part of the hospital to another part of the same hospital, unless someone literally walks it there. The fact that, you know, you can do stuff with AI. I've had clinicians, I had my liver surgeon, was massively open to it. And he was sending me, he was asking me for updates using my tool.
[00:12:01] And then he was sending me videos of him going over my scans on WhatsApp because it was quicker than, it was quicker than using the internal systems. He could send me that. Super safe. Yeah, yeah. Super secure, super safe. And I just think if you've got professionals that are open to it, that's really important. And I think my sister's a GP and she's, I know they're using some tool internally now to help them with like note taking.
[00:12:28] And so I think it will get there, but I think you can't rely on, the NHS is going to take years to catch up. Even private healthcare is going to take a long time. So I figure the more control you can take yourself, the better, really. Yeah, yeah. You mentioned that you use Claude for symptom tracking, LLM for just long term making sense of the data and browsing through it.
[00:12:55] And then you also keep a hard copy in Google Drive or Google Docs. How did you decide which tool you're going to use for what? Obviously, the hard copy is self-evident one. But when it comes to LLMs, did you ever compare the outputs the different LLMs give you? How did you decide where to go and what to stick with? Yeah. I just through trial and error, really. And I think a lot of it's personal.
[00:13:23] I think I like the way I was stuck with ChatGPT for ages. I tried to move to Gemini because my work was used to Google Workspace. So Gemini is improving all the time. If you read into any of them, they all release new products constantly and they constantly jump each other in terms of perhaps how effective they are. I never used Claude, partly because I think I was stuck.
[00:13:49] Yeah, partly because I was stuck between ChatGPT and Gemini. I also tried there's a service I mentioned before called Manus, which I think has been bought by Meta. So it's owned by Facebook now. So that put me off them a bit. And then I just started using Claude really recently, probably in the last two months. And I did it because I was reading about how it's more AIs is generally built to agree with the person that uses it.
[00:14:17] So a lot of them don't really question a lot. They will. One of the problems a lot of AIs have is that they'll back up what you say and they'll often not be overly critical. And Claude seemingly is one of the better at being a bit more critical where it needs to be. Being a bit more, I just didn't, I like the way it responds to how you use it. It's not perfect still. And I don't think any of them are, but there's more and more people using Claude now.
[00:14:45] The functionality on it is incredible with if anyone had a chance, they run, they have this service called Cowork, which allows it to take over your system. So it's, it goes one step further now, I think than any of the other tools that I've used. And yeah, it's just reliable. I use a couple of other tools with it. There's a, there's an amazing tool called Whisper Flow, which is great for talking when I'm feeling rough, especially when you just need to talk into an app.
[00:15:13] Whisper Flow sits across like WhatsApp, sits across anything, can be used in Gmail, can be used in Claude. You just talk. It's got amazing capacity for understanding what you're saying and correcting it and making it grammatically correct. Which a lot of the chatbots, a lot of the chatbots you can talk to, but their audio sort of recognition isn't very good. So Whisper Flow is amazing for helping with that. And yeah, and that's really, and just trial and error, really.
[00:15:38] I think, as I said before, you really need to be paying for them to get the full breadth of what they can do. The free tools aren't really, the free tools are fine for a trial. But yeah, if you're going to throw data, private data into them, I would, I wouldn't do that. And the other thing I use, the other sort of daily thing I use Claude for is I take a lot of supplements. I take loads of pills, basically. Some from recommended by my oncologist, some recommended by a nutritionist. I inject mistletoe. I do loads of stuff.
[00:16:08] And keeping track of that is a nightmare. So I have one tool that's just built to tell me what to put in my pill pots every two weeks. Sorry, every week. And that sits separately from the symptom tracker because I think you should never try and get one tool to do too many things. If you've got like distinct uses, it's, I'd say it's better to have separate tools.
[00:16:32] Like I've made the mistake before of trying to buy, trying to do like single tools that do loads and they mess up because they just, it gets too complex. Yeah, yeah. So that's amazing. That's, that's, that's just the thing. So how do you make sure that you always have the critical perspective and the critical mind when you are posing the questions to the LLMs? They do have confirmation bias. Yeah.
[00:17:02] So you need to test them out a little bit. I started comparing different tools with the same prompts or just for tracking my own symptoms as a chronic patient. And I noticed that Gemini seems to be the most wannabe comforting and you're awesome type one. Like you would say what you're noticing and it would be like, oh, that's super introspective of you.
[00:17:30] And you're like, no, it's not. Yeah. And I agree with you. Like I was, I never used for the same reason, stuck on ChatGPT. But ChatGPT recently, I don't like it because it lost the narrative. And now it's just like doing too much of bullet point, very short responses. And Claude, I was blown away by the way that it structures data, the way that it can create summaries for you.
[00:17:56] And also it has a little bit of a in the middle type of and trying to please you approach. It's not overly excited about you and everything that you're doing. It just gives you a lot of useful information. So what kind of prompts do you use and what's constantly in your head that you don't go in the wrong direction?
[00:18:22] For example, I always try to think, don't ask a yes and no question. Always do what and how. How does this work? What does this mean? Instead of does this mean that I have blah, blah, blah, blah, type of a thing. So what's your tell me a little bit more about your mindset when you are asking questions and when you are trying to make sure that you stay critical? Yeah, I think it varies. I think a lot of the time.
[00:18:50] Yeah, if I'm not, I think it's a really good question. I think I would use words like I would ask it to be I would ask it to be critical and ask it to question one. Whenever I build a lot of the Claude stuff, I run in projects which operate, you can give them generic rules about how you want them to behave. So you don't have to, you know, hopefully you don't have to keep telling it the same stuff.
[00:19:14] So whenever I create a project or like a custom tool, I would always have standard stuff in there about being critical, being but avoiding bias, being making sure that it's just telling it that you telling it that you want it to question you and you want it to check and you want it to check the information that you're giving to make sure that it's accurate and you're not misleading the tool.
[00:19:39] And I think with general prompts, like you say, it's not you don't want to I don't know what the word is, but like you that example you gave is a good one where you can in some ways you can give it the answer. You almost can accidentally give it the answer that it feeds back to you. It's like that confirmation bias. If you're too big on one, if you're too big in terms of how you represent something, the likelihood is if it thinks it wants to hear something from you, it will present that back to you. And I just think you've just got to be really careful about how you ask questions.
[00:20:07] Yeah. And if you've got any doubt, ask it to avoid bias, be critical of what you're what you're saying and ask any questions that it needs to ask you to ensure that it comes back to you with a balanced answer as opposed to as opposed to just confirming whatever you think it needs to ask. But I do think like you say, I do think Claude is a bit better at it. And the more with any of them, the more you use the same tool. I think that was why I got stuck with ChatGPT for so long, because it does get to know you.
[00:20:35] It gets to understand who you are and what you do. But Claude seems better at that than any. Or constantly you can go into the settings and you can see how it's building a profile of you, which you can update. If you want to put like consistent rules into it, like, you know, about things like bias, you can. And I think you just have to be wary that sometimes there are really important, like, for instance, tracking my symptoms. I'm not that worried. If it says, oh, you've had a great day, blah, blah, blah.
[00:21:04] And I felt and you're like, actually, I feel pretty shit, but the numbers might look all right. But I wouldn't necessarily go back and correct it. Whereas, for instance, I've got a like co-work, Claude co-work set up, which is essentially just, I call it my cancer smasher weekly.
[00:21:23] And basically that's got over the years, I've done so many tests, both through the hospital, but also I've done paid for like DNA related tests on my cancer profile and my tumor, which I've done through companies like Astron Health and others. And I've got loads of data that I'm like, I was like, what on earth do I do with this stuff? I don't understand half this stuff. All my oncologists have seen it and they're probably not going to do anything else with it.
[00:21:52] So I plugged it all into this Claude tool and I regularly update that with my bloods. And that I've trained to just do every week. It will search across the latest cancer research, the latest trials. It'll cross-reference it with my humor type as best as it knows from all the knowledge that it's got. And it'll flag anything that might be relevant for me to read. And I think things like that, I'm like, it's incredible what it comes up with.
[00:22:20] And sometimes it will come up with things that are raised or I'll think, I'm not raising that with the oncologist because it's a trial somewhere that's never going to be relevant or it's too early days. But every now and then it throws something up and you just think, actually, who's doing that? Is anyone else doing that for me? Probably not because not many people have that level of information. And I think that's where you want to make sure you're avoiding too much bias.
[00:22:47] But if it's got all the data and you've guided it well with prompts, I think it's only going to, I just think it's going to, it's got to do more harm. No, sorry, other way around. It's got to do more good than it does harm. I can't see many things going that wrong. So yeah, that's my, and I think that's my philosophy really. I've, as I said before, thought about, I just think there's so, you just think, don't you? Like when you've got like yourself, there's so much out there.
[00:23:14] There's so much AI has capacity to solve so many things. It will solve. I'm fairly confident that in, I don't know how long, there's got to be cancers that will be cured as a result of the work that AI is what people were using AI and combining it with medical research. Yeah. I just think it's a matter of time. And I don't think I'm going to do it, but people are doing things all over the world and it is, it's got to be worth a try, right? Yeah, absolutely.
[00:23:39] So what do you observe that using AI is doing for you just psychologically in terms of, I don't know, managing fear, managing anxiety? Is it increasing anxiety? Because I would put patients in three buckets. If you're not careful, you can go down the road.
[00:24:06] Like the best case scenario is you become a great partner to clinicians. You even come up with new ideas. They're open to taking these ideas into account and you get better outcomes. Then there's the scenario when you just are trying to calm yourself down. So you go down the road of minimizing the problems. And that's the research that I did with just deep research around how patients are using AI.
[00:24:31] And it often happens that they would not go to when they should. Yeah. Because you're like, nobody really wants to go to ER unless you like really are in horrendous pain. And then there's the third option where you go after the worst case scenario.
[00:24:54] You go down the rabbit hole of the worst case scenario and you are potentially overburdening the system. So what has been the case in your case? I don't know to which you're thinking about this. Yeah, it's a really good question. I think it's really funny because I, ever since the beginning, I literally remember going for a walk with my wife when we found out on the first day that I probably found out I had bowel cancer.
[00:25:21] And I put it into Google once and never did it again. And actually, I realized almost straight away that my, that wasn't the way that I wanted to read about anything. I just think, I think, and I think across the board, I think I've said, I've taken the same. I don't know if my approach is the right approach. I don't really read anything about it.
[00:25:46] I'll listen to doctors and I'll listen to nurses, but I hate any booklets, any pamphlets, most websites about cancer. I've avoided like the plague. I have spoken to charities, spoken to people. I think speaking is the best thing I've done. And so, yeah. So when it came, when it comes to AI, I think I've used it in a similar way. I haven't used it to ask questions about my condition and things I should do.
[00:26:13] I don't put in, I might have put in, if I've got a problem with my skin or something like that, something like a minor, a minor ailment related to maybe something that I'm doing. But I wouldn't ask it anything serious. I'd always go to my nursing team. I think I would always, I would never, equally, I've never gone to, I've never dialed 111 or called 999 or gone to A&E with anything. I've always just spoken to my nursing team.
[00:26:40] And I think as a result of that, I've used AI to make me feel like I've got an element of control over something that, you know, deep down, I've got very little control over. Yeah. But there are a few things that you can, there are obviously a few things you can control.
[00:27:01] And if you get, if you get things like simple things, like when you're on chemo, if you, I've found that if I eat at the right times, if I sleep at the right times, if I exercise at the right times, if I socialize at the right times, it's all pattern based. And if I get that, it makes a massive difference to my day. Yeah. And I think that's how I use it.
[00:27:24] It's the, it's just giving you, funnily enough, one of the big things I did was launched when I started my website. I launched it on LinkedIn. And I did a kickoff post where I had just put something like, I think it was like, fuck you, cancer or something like that in a big message. And I had no, and I'd realized that I'd never really said anything to everyone. I had a very close group that knew what was going on, but I had a lot of wider connections.
[00:27:54] And I realized that I felt like it was holding me back a bit because you don't want it to define you, do you? But you do, it plays a big part in how you think about things and does change your mindset. And it does a lot of things, which I think probably have a positive impact over time. It makes you more focused on the things that you probably should be focused on and less focused on all the other shit that you worry about that you realize actually isn't important. Yeah.
[00:28:19] And I think the AI thing, the combination of coming out on LinkedIn, starting to blog about it on my website and being honest about what cancer is like, what chemo is like, being clear that life carries on. And there's no reason, granted, at times it can't carry on and you've got to just lie down and try and get over it. But yeah, it's given me a massively, I feel, whether I should or not, I do feel really positive about what can happen.
[00:28:49] And yeah, I just think, I just think it helps. Yeah. If you can take some control, if you can, if it makes anything that you can do to make yourself feel like you're being proactive and you're not just waiting for the next bad thing to happen is a massive bonus. And I feel like that's what AI, among other things, has allowed me to do really. Yeah. Yeah. I agree. And I like the fact that you mentioned the control part.
[00:29:15] It's a huge part of being a patient, also being a chronic patient where, you know, sometimes you won't do something that technically speaking you might supposed to do just because of wanting to have control. But I want to share another example of just the mind shift that happened with me around control last year. So I've got IBD.
[00:29:41] It's, I've had it for 23 years now and had some problems last year. So obviously, you try to do so much stuff. You think, oh, maybe I'm too stressed. You try to de-stress. And then like when things don't improve, you're blaming yourself that you're not doing enough of the things or that you're doing something wrong. And then I said, I know you shouldn't really use JGBT for as a therapist. And I do see a therapist. And I do see a therapist. But I was like, let's see what happens. Let's just explore.
[00:30:11] Let's see what people do. And I went to it with a critical mind. But what was super interesting to me was that it said, can you, what if you took the scenario of knowing that biology is complex and we don't understand everything. And maybe you do not have control and maybe you shouldn't blame yourself because people want control.
[00:30:35] And even when you blame yourself, that means that you are asserting control because you live under this impression that if you just did something differently, you would be okay. So at that point, I was like, oh, maybe maybe I shouldn't blame myself for what's happening to me. And that made like a massive difference in terms of how I felt. And that doesn't mean that I suddenly started living on the edge and stopped caring about what I do.
[00:31:04] It's just, hey, maybe it's not my fault. And it's also the way that medicine also sometimes defines what happens in the medical journey, like patient failed the treatment. I have a Canadian friend who says, who's super focused on the language used and she's patient didn't fail the treatment. The treatment failed the patient. So super interesting stuff when you do that. Yeah, no, it's absolutely true. I think it is really individual.
[00:31:34] But yeah, I think there's lots of, a lot of it's a mindset thing, isn't it? And just changing how you, trying to change how you think about things. And I did, I think like most people do, when I initially found out, it hit me for six. And I, and also, I had loads of, I had loads of, I had surgery and stuff and put my body went through loads of stuff. Awful, like really hard things that, and it's hard when you're mine. I remember being in hospital.
[00:32:01] I woke up in hospital on my 40th birthday, having had like surgery on my lungs, I think, or liver. I just remember being like agony. And I was just like, I cannot do this. I had definitely, that was my lowest point in hospital. And I think you've just got to, you come out the other side. And like right now, I do feel things have moved on. And to be fair, my cancer is way more serious really now than it was then.
[00:32:30] But yeah, you've just got to take the good days, right? And just do the most, I think just do the most you can with them. And like you said, use, I think people should try AI for everything and anything. I do think the counselling thing is a really interesting one. I actually had some counselling early on, which I don't have anymore. To be honest, I found putting stuff on LinkedIn almost like better than counselling.
[00:32:53] Because I think I just found it a place that I could talk about things with an audience, which, I don't know, it's weird. I'm not on Facebook or anything. I don't know if I'd have done it on Facebook. And maybe LinkedIn isn't the right place to do it. Maybe it's exactly the right place to do it. Yeah, I can't decide. It definitely helped my relationship with work. It helped me balance. It made me realise there were so many people that I worked with or used to work with that didn't know. And it opens it up.
[00:33:22] And you have this, I have now this really weird combination, I think. I think cancer has made me maybe narrow my friendship group. Because you cut the wheat from the chaff a little bit and you focus on people you really want to be friends with. And you really want to keep up to date. So I've definitely chopped out friends, not intentionally, but just have a more friend. Really good friends now. But then I have this weird LinkedIn profile, which is very open and large and keeps things. Are you me interesting?
[00:33:51] I wouldn't have met you, obviously, without LinkedIn. I wouldn't have met Dale. Dale, there's loads of other people that have been in touch on LinkedIn that has been fascinating. And they get in touch and ask what I'm doing. And they explain what they're doing. And opportunities come out of it. Yeah, it's been weird. It's been... And those two things, the combination of those... And AI has helped me across them. AI has helped me with getting stuff on LinkedIn. And when I'm in hospital, just recording messages.
[00:34:19] And, you know, people wang on about messages being written by AI. But I think if you're using it as a means to talk and get stuff out there, I do it. Because it's easier to talk into an AI tool and get it to write something that you can then edit. Yeah. Don't be able to type it if I'm in a hospital bed or whatever. How do you notice any differences? Like models, AI models are advancing really nicely.
[00:34:48] It's amazing the progress that they're making. Did you notice any differences? Or how did the advancements of AI models impact the outputs that you're getting or all the tracking that you are doing? Because you have been doing this for quite a while now. Your patient journey, the way that it develops, is also going through this whole AI transformation. And I'm wondering to which degree do you see that? Or what impact does it have?
[00:35:15] I think the main thing I've noticed is I think the biggest developments I've felt are things like move towards some of the tools basically becoming properly agentic. And like Claude Cowork being the best example of that. I haven't actually used, I tried to use Chatty BT's sort of version of it, but didn't really get on with it. Gemini hasn't really opened it up yet. But Claude Cowork provides you with this amazing tool to, you can set up.
[00:35:45] At the moment I've got, I don't know, seven or eight different co-works that all do different things for me. Some related to work. Some related to other things I've got going on. Some related to cancer. Then I use my phone. If I leave my laptop on somewhere, like at home, I can take my phone with me. Claude's got a service called Dispatch, which I don't know if you've come across. It was like announced in the last two weeks.
[00:36:09] And Dispatch is just a single screen where I can interact with all of those co-works back on my laptop. So I can tell you can almost use Dispatch like a, I don't know, it's like having a managing director of your business really. And then the co-works that you've set up almost act like separate employees. So you can say to the co-works, you can go into Dispatch and say, I need you to, there's a problem with the web. There's a problem with my F cancer with AI website.
[00:36:37] I need you to sort out the X or I need you to write a new blog about Y. And I need you to help me with a LinkedIn post for it. And I need you to, my symptom tracker's off. I need to track, I need a quick update on what happens to me every fourth day of my cycle. And Dispatch will go away, work on all those tools, get all those jobs done separately and report back to you on one screen on your phone.
[00:37:07] And that is a game changer because I can go to hospital now and I used to take my laptop. Now I just take my phone, leave my laptop at home and just count on the fact that it can do it all and report back to me. And then I can read, if I want to read long documents and stuff, I can do that when I'm at my computer. So that's made a huge difference. It's also made a huge difference on how much money I spend on AI because it is a lot of computing. So my core subscription is up to, I don't know, £200 a month, which is quite a lot.
[00:37:35] When it was, it was £20. But it's meant, my efficiency, I think is like through the roof. And I suppose the other big change I've noticed is I love Notebook LM, Google's tool. It's free to, I think it's free to everyone. Obviously, you are a paid Gemini customer, you again have the security thing, which is a bit better. But it's more secure anyway because of how Notebook LM works.
[00:38:03] And the stuff you can do with that, you can dump loads of research. If there's a new cancer paper or papers that you should read and you're like, or even a book and you're like, Jesus, I'm never going to read that. You can chuck it in there and it can turn it into a podcast for you. It can turn it into an infographic for you. The stuff that Gemini and Notebook LM can do with images is amazing.
[00:38:30] And I use that for, I took out my, the other day I took three, sorry, six weeks of my symptom tracking and put it into that. And then it turned it into this amazing infographic that I just stuck on LinkedIn. It also turned it into a 12-deck PowerPoint slide, which I thought was a bit excessive for LinkedIn.
[00:38:51] But amazing for me to look through and be like, oh Christ, there's a lot more to just, there's loads of things I haven't thought about that it's captured in here. Yeah. Yeah, amazing stuff. And the other thing, slightly random, I found I put in a load of, I put in all my doctor's notes into Notebook LM and all my blood results. So I can always refer back to them. And the good thing about Notebook LM is it's a closed environment.
[00:39:18] So it doesn't, unless you tell it to look at a website or something, it doesn't. It just focuses on the information that you give it. And I got it to turn it into an audio file about me and my, like, I hate the word cancer journey, but that, but cancer journey. And it, and within it, I was listening to it just in the background. So I was like, oh, this would be quite funny. I've got no idea what it's going to say. And I found out when listening to it that I'd had my gallbladder removed and no one had ever told me. It just happened as one of, put in one of the surgeries that I had. And the Notebook LM.
[00:39:47] That's a bit of a big thing. Notebook LM. Did you, did you then check if that was true? Yeah, it was true. It was true. You didn't know? No, I didn't know. It happened as part of my, I can't remember, stomach surgery or liver surgery. They took my gallbladder out. And I guess because it was minor and you don't need it. And the surgery was quite a big deal. No one thought to, no one thought to tell me it, but they just stuck it in my notes. Yeah, that's how I found out. So that was quite weird.
[00:40:16] That is questionable on so many levels. But okay, let's not, let's not get into that too much. Did you, did, did you ever learn anything that turned out to be false? Did it ever happen that you would go in the wrong direction or that you got the advice that it wouldn't be correct? No, I don't think it has actually.
[00:40:43] I think probably, I probably am fairly realistic and don't take anything with. I think you have to, if you're reading anything from AI, you've always got to question it. A bit like if you put anything into Google, you've got to question it. I just think probably you're going to get, I'd say you'll probably get a more accurate response and a tailored response in AI than you're ever going to get when searching stuff on Google.
[00:41:08] And as long as you just keep, remain questioning, don't believe in everything you read. There's so much rubbish about cancer on the internet. I just think AI is a good place to start. It's also great for, I, for questions you've got, if you see a story that you think you're not sure about. I've had loads of my, I've had people in my life, lovely people send me stuff and they're like, oh, you should read this from this guy who's in India and he's trying this new
[00:41:38] thing. And you're like, okay, it's not worth looking at. But I had a very close member of my family did that a lot and meant, and meant well. And I actually, it remains to be like, look, you might want to put this into AI and actually ask if it's legit and what it is and what it's about before you send me some of this stuff. I don't think I've ever, I think you probably could go down a rabbit hole with anything, but as long as you question it and as long as, if you use AI to question, you know, to question things like that, I think it's quite a good approach.
[00:42:09] You've been listening to Faces of Digital Health, a proud member of the Health Podcast Network. If you enjoyed the show, do leave a rating or a review wherever you get your podcast, subscribe to the show or follow us on LinkedIn. Additionally, check out our newsletter. You can find it at fodh.substack.com. That's fodh.substack.com. Stay tuned.