AI isn’t taking jobs — it’s changing what jobs are. On today’s episode of the Me, Myself, and AI podcast, GeekWire’s Todd Bishop joins host Sam Ransbotham to dive into how artificial intelligence is reshaping work, learning, and creativity — not by replacing humans but by amplifying what we can do. From classrooms where students use AI on exams to newsrooms rethinking how news stories get written, they explore the opportunities (and headaches) of this new era. It’s a smart, funny, and refreshingly real look at how we’re all learning to work with our newest coworker — artificial intelligence.
Subscribe to Me, Myself, and AI on Apple Podcasts or Spotify.
Transcript
Allison Ryder: AI tools are evolving at a rapid pace — so quickly, even, that they change when users go back to them. On today’s episode, we speak with a journalist and fellow podcaster about tech advancements and their implications on organizations.
Todd Bishop: I’m Todd Bishop from GeekWire, and you’re listening to Me, Myself, and AI.
Sam Ransbotham: Welcome to Me, Myself, and AI, a podcast from MIT Sloan Management Review exploring the future of artificial intelligence. I’m Sam Ransbotham, professor of analytics at Boston College. I’ve been researching data, analytics, and AI at MIT SMR since 2014, with research articles, annual industry reports, case studies, and now 12 seasons of podcast episodes. In each episode, corporate leaders, cutting-edge researchers, and AI policy makers join us to break down what separates AI hype from AI success.
Thanks everyone for joining us again. Our guest today is Todd Bishop. He’s the cofounder of GeekWire. Todd [has] a deep background in journalism, starting back in the day, I think, [with] student newspapers, but more recently, founding GeekWire back in 2011. Todd, let’s start there: Give us a little background about GeekWire.
Todd Bishop: Well, my colleague John Cook and I were longtime newspaper reporters, and we saw the blogging revolution happening back in … 2007, 2008. And from that moment on, John has kind of been my coconspirator and somebody who’s inspired me to really push the envelope with embracing technology in reporting. We were writing for a daily paper in Seattle called the Seattle Post-Intelligencer, which no longer exists in part, I theorize, because the name is so dang hard to pronounce.
Sam Ransbotham: As evidenced.
Todd Bishop: So we, through a variety of different twists and turns, ended up starting GeekWire in 2011. We’re based in Seattle, and we like to say that what happens in Seattle matters everywhere. The whole notion is that there’s a worldwide audience of people and companies and all sorts of things that basically want to know what’s happening in the Pacific Northwest. And that is our niche. GeekWire is a small media company, events company.
I have been really thriving, from my perspective, over the past two years in this new AI world, based on two things. One is the much more interesting landscape that we have to cover and the extremely powerful tools that we suddenly have at our disposal to use in the process of reporting.
Sam Ransbotham: Well, you said it’d become very interesting, and I think that’s a nice euphemism. It’s a little chaotic at times, isn’t it? Just for background to our listeners: Todd [has] a podcast. What are you seeing [that] is different about the world from when you started this GeekWire approach and the podcast to now?
Todd Bishop: I would say there’s one thing that really stands out — obviously there [are] many different things that stand out — in terms of the illustration of how different the tech world is today.
We have a feature called the GeekWire 200, which is a ranking of Pacific Northwest startups. From the very beginning, our core metric was employee growth. The question was: Are you growing? Are you raising money? And are you adding people? And if so, that was our barometer of success. That has just been flipped on its head, to the point where it’s almost the inverse, where if you’re growing and succeeding with fewer people, frankly you’re probably more successful, at least in terms of reducing your cash burn and doing more with your money. To me, amid all the change — we could sit here for an hour talking about everything that’s changed — is the clearest illustration of how different the world is today.
Sam Ransbotham: Actually, I hadn’t thought we’d get into this part so quickly, but you brought up the specter of jobs. We’ve seen so much talk about jobs and how much … AI is going to take over all the jobs. I’m not seeing it. Are you?
Todd Bishop: I’m seeing AI take over tasks, not jobs, and that, to me, is the big difference. In that way I think it’s emboldening executives, especially [at] these larger companies, to let people go, to assume that the people who are left will be able to do more with the AI tools than the large numbers of employees would’ve been able to do without them. So in that way, it’s an indirect impact on jobs, but it’s very hard for me as a reporter to point to a job, especially in the tech industry, where it’s one-for-one: AI is now doing this particular job, and therefore that job no longer exists. It’s a little more indirect, but what are you seeing?
Sam Ransbotham: I think I push the opposite way. Actually, I was listening to one of your recent episodes with the Box CEO, and that really, I think, aligned with what I was thinking too. I’m really firmly embracing this idea that this is going to be a massive job creator. And I know that’s sort of counter to the “Woe is me. …” Like you pointed to a second ago, the perhaps nefarious mode is for announcing job cuts, which may or may not be related to a task. But the ability for us to do more than we’ve ever done before is going to create opportunities.
Todd Bishop: I’ve brought this up so many times. You mentioned the Box CEO [episode]. I think the GeekWire podcast listeners who might be tuning into this are going to be sick of hearing it, but I’ll mention it.
I went to Aaron Levie, the Box CEO, and I assumed that he would be using AI agents to basically do less work, right? He’d be off in Tahoe skiing, water or snow, whatever he might be doing that season.
Sam Ransbotham: Whatever phase date.
Todd Bishop: Exactly. It was just the opposite. This was a really eye-opening moment for me. It was about six months ago [when] I talked with him. The fact that the horizons have been expanded so much for him, for the company, there’s so much more that can be done now with AI, and, therefore, there’s more time and energy being put into doing those things. It’s a little bit the opposite. To your point, I think when you think about the job market, there’s really the potential for all these new possibilities to create more, not less. Granted, there’s going to be a painful transition here in the meantime. But I think in the end, I tend to buy the argument that this is going to be additive, much like prior revolutions in tech and industry.
Sam Ransbotham: I think about it from a value perspective: If something costs less, we consume more.
So what these technologies are doing is allowing us to create some components of technology — not all components, but some components — for less, which says prices will drop. And if I remember my p’s and q’s right from [microeconomics], that’s going to play out with a greater quantity, but, like you say, I don’t want to be cavalier — [they’re] not going to be exactly the same jobs, [they’re] not going to be exactly the same tasks. We don’t have a lot of travel agents like we used to have prior to the internet. One of the BLS statistics that I remember was that travel agents were forecasted to be a massive growth opportunity because of the argument that increased discretionary income would lead to more ability to travel, [and] more flexible jobs would create demand for more travel agents. As we know, that didn’t exactly work out that way.
The mechanisms were right, but we tend to think so myopically when we’re projecting. We project along one dimension, and then, well, everything else changes too.
Todd Bishop: Totally. I’ll turn it back on you a little bit, Sam, because I’ve been listening to your podcast now, and one of the ones that I really appreciated was Jeetu Patel from Cisco. One of the things you got into with him was this whole notion of the generational differences and the generational viewpoints, and the idea of not letting this be an opportunity to no longer have interns or entry-level employees, but to, in fact, look at this as a chance for almost co-training, where the people who are coming into the workforce and taking these jobs, these entry-level jobs, are just as much training the people who are further along in their careers — talk about a euphemism, like you and me — as the people who are further along are training the newcomers.
That really was one of those aha moments. It was actually on a level of what I learned in terms of Aaron Levie telling me, “No, this is more work, not less.” It’s like, no, we actually need to create more interactions with these people taking these new jobs and not just rely on the AI to do all these things.
I’ve had a couple of different interactions over the last three months with entry-level reporters who are just coming out of college. The dynamic is really interesting because they’re just coming out of school. Maybe you can address this in terms of some of your students, Sam. In each case, it was an intern and then somebody else I was mentoring who was going for his first reporting job.
I said, “How much are you using AI?” Their first reaction was to get defensive and think that I was going to accuse them of using AI. And I was actually just thinking about it in, obviously, the complete opposite [way] like, “Are you learning the tools? Are you figuring out how you’re going to fit into this new environment that the world is turning into, and what kinds of ways you’re going to use them responsibly as journalists?”
So it’s interesting, I think, in some ways from my vicarious observation there, my view into the world of academia from somebody who’s just seeing these people come out of it, I wonder if the messages are right. And I’d be really curious for your take on that from your own firsthand experience.
Sam Ransbotham: Journalism and education are two — I guess everyone thinks that their own industry is particularly subject to any new technological revolution. But certainly journalism and education feel at the forefront of some of this. I liked the way you asked that question before with one of your guests: Is it cheating to use AI?
But one of the things that I do in our class is we are wide open, [we] use AI for anything, even on the exams.
Todd Bishop: Oh wow.
Sam Ransbotham: The idea there is that, all right, this is the new tool, and we have a whole set of other tools we’ve expected you to learn how to use. You need to learn how to use this one too.
Now, I’m not so naive as to think that there are no classes that doesn’t work for. I think the classic example is if we learn how to add and subtract, we don’t want people using a calculator. That’s now become a trope. But if you’re not doing the adding and subtracting classes, then the calculator is a pretty useful tool. And particularly in terms of journalism, everybody’s going to be using those tools.
I don’t think we’re saying, “Don’t use these tools” though. I mean, maybe don’t use a chainsaw in a garden party; that may be a pretty good rule of thumb.
But people are going to be using these tools in organizations, and they’re going to use them before we have infrastructure in place. And that’s just kind of the way things work. We’ve got no other human civilizations to point to and look at the back of the book and say, “Well, here’s how that last human civilization did with artificial intelligence. This is what we should do.” We all just kind of have to figure it out.
Todd Bishop: [I] totally agree. I do like the idea of thinking about AI as a power tool. You don’t just hand it to somebody who’s never really used it or thought about the implications. Maybe in a personal situation where you’re just trying it out and testing it. But if you’re truly going to be using this in a company … I think it’s really imperative on companies to train, but even more so on individuals to figure out how to use it through some trial and error in low-risk situations. And to me, that’s really important.
Sam Ransbotham: For the record, I did not pay Todd to say that. I love that analogy, though, because I can go to my big-box, do-it-yourself store, and I can buy any of these tools, but what I do with them is very different than what someone who actually knows what they’re doing with that tool would do.
And I think about that with vibe coding, which I’m still so uncertain about. I love the idea that we’ve lowered the barrier, that people can use these tools; they can create things on their own. At the same time, to what degree … I’m not really ready for air traffic control to be working off somebody’s vibe code.
You’ve been doing this podcast for a while. What do you think people are interested in hearing about? What makes a story interesting for people? What captures the zeitgeist?
Todd Bishop: Things are moving so fast that newsy episodes tend to do very well for us. Just recently, I got a chance to talk with Ed Bott, who was a longtime journalist who’s covered Microsoft. We talked about the latest AI innovations in Windows, and when you combine AI with something that’s as widely used as Windows, and the whole notion there, in particular, that people will be able to use agents on their own files, on their own desktop — that kind of mix of timeliness and utility.
Another one that really did well for us recently was a discussion about vibe coding with Chris Pirillo, who’s a longtime geek and conference organizer.
Sam Ransbotham: You talk about news stories doing well. One of … our challenges, though, is to pull away from some of that, and I probably say that myopically, because we are trying for more enduring types of themes on our show than you’re driven a little bit more by the news cycle. But you know what, if we’re chasing all these things, I feel like it’s really easy to get super invested in tool A and then find out that tool A changes on you. I think you had an analogy, I was listening to one of your episodes, that something about one of the tools changed out from under you. You used to have it fact-checked for you, and then it quit fact-checking.
Todd Bishop: Yes.
Sam Ransbotham: Am I remembering that right?
Todd Bishop: That’s right. It was Gemini. It used to be that it knew implicitly what I was doing when I uploaded a story and related materials to fact-check. And it’s really interesting; it kind of goes along with the vibe coding theme, where things are shifting and changing.
What happened was I uploaded the story and the underlying materials to Gemini and gave it my typical prompt. It was something like, “Hey, check this story,” and it used to just know what I meant. And suddenly it was doing something really weird. I can’t even remember exactly what it was doing, but it was kind of pretending that it was fiction and giving me feedback. … And that has been a challenge in some respects because these are not human-created products in the end that we’re using. It’s not something where a product manager is saying, “And when a user does this, the UI will say this.”
I think it’s a challenge not only for the users like us but for the companies that are making them. I think in terms of big-picture trends, and I agree with you, we do try to find some deeper insights in the day’s news or the week’s news, and that is one of them that I think is a completely different kind of product that we’re dealing with here.
Sam Ransbotham: I had my own personal experience where I created some course content with the assistance of AI and, of course, only got halfway done before the semester started. And so then I went back once to catch up, [and] it completely had changed how it was working.
If you and I are dealing with this at the individual level, what’s it like for enterprises to build systems that are, in theory, working with these tools? If they don’t use them, they’re screwed. If they do use them, they’re screwed. Is there a path in here that’s not “you’re screwed” at the end of it?
Todd Bishop: This gets into your Hugging Face conversation, and I know we’re coming off like superfans of each other’s episodes.
But we have been listening to each other’s podcasts here, which is good podcast host practice, I think, before you’re going to have a conversation with somebody. I think one answer is open source. The fact that a lot of these models are available for companies with sophisticated developers to bring in and have a modicum of transparency, and maybe not control over the outcome, but at least the ability to put some parameters around it and to control the environment.
We ended up covering a nonprofit in Seattle called the Allen Institute for AI. I go in — it’s kind of cool, this used to happen more in tech coverage, where you would go in for a product briefing, and you would walk out like an hour later with not only information about the thing they were going to be releasing but also just a much deeper understanding. And anytime I go to Ai2, the Allen Institute for AI, I walk away with … like, “Oh, now I understand what post-training AI is.” It’s that kind of institution.
But more to the point, I do think when you’re talking about control and the ability for companies to have some predictability in the way they roll out these AI models and what the AI models do, I think open source is one of the answers.
Sam Ransbotham: One of the things that I’m thinking about now is that we’ve focused on these tools in their ability to generate things, maybe myopically ignoring the demands that they’re putting on consumption, meaning that we are now able to have far more content than our ability to consume these. And that’s true in organizations. That there’s a whole bunch of organizational stuff that has to happen too. That just because a tool is better doesn’t mean we should put it in place, because it may not be enough better, or it may be a little fragile. These are all consumption-oriented problems, less than production. Can we do anything about the consumption problem?
Todd Bishop: That’s a great question. I tend to think of this in the category of “let your bot talk to my bot.”
Because, as a reporter, and of course everybody has this problem, I feel like I’m getting more and more email, in particular, coming in, just overwhelming me. I think it’s getting worse, in the same way that spam calls are getting worse, in part because people are using AI to optimize the content or to at least — optimize is probably the wrong word — accelerate and expand the content, to your point.
My solution to that, in part, is AI tools. Maybe AI can make our communication better.
Sam Ransbotham: There’s some hope there. You mentioned all the emails you’re getting. We get a ton of emails every day. It’s really quite phenomenal how much, and I’m sure you even get more, but there [are] some great emails.
How do you engage with listeners from a podcast standpoint? One of the things that I personally find frustrating is that you and I talk to each other, and it’s fun, but I’d like to talk with everyone else, yet all these emails come in. How can we better engage with listeners, as podcasters?
Todd Bishop: This might be sort of a call to action for me, Sam, because I’ve thought about this a lot. And, if anything, I think of it even beyond listeners but also readers as well, because we tend to be text-on-the page first, and then the podcast is sort of after that. And I would say, about six or seven years ago, our comment section just went off the rails. And it was no longer adding value other than some kind of malicious gratification that somebody was getting from insulting somebody else who had posted something else.
Sam Ransbotham: Venting ground.
Todd Bishop: You know the deal.
We’ve talked about bringing Bluesky comments back in using their open protocol, and that’s something we might still consider. But I miss that one-on-one, direct kind of interaction. In terms of podcast listeners, I’ve got, I’d say, about 10 to 15 people who are the power listeners.
Sam Ransbotham: Exactly.
Todd Bishop: If it’s a good episode, I know that I’m going to hear from at least a handful of them. And I’m sure you’ve got the same thing where it’s like, “I thought about this and what about this?”
Sam Ransbotham: “Well, you got that wrong.” [Laughs.]
Todd Bishop: Yes, exactly. “You got that wrong. Why didn’t you ask this?”
Sam Ransbotham: That’s my favorite one. You know, usually they’re right. I should have asked that. Shoot, what was I thinking?
Todd Bishop: Exactly. But we haven’t figured out a good way to systematize that and to incorporate it back into the show. And part of this is just time and energy and resources to do that.
Sam Ransbotham: I think about it, coming back to you’re in journalism, I’m in education. We used to have a model in education. They call it the banking model of education, where empty students come in, and the learned professor dumps a bunch of information on them, and they walk away filled with knowledge. I don’t know if that ever worked, but it certainly doesn’t work anymore.
Education seems to work when it’s bidirectional, when it’s colearning, when there’s a benefit to being in the same room, engaging with people. Much of the journalism world doesn’t support that well because of exactly the problem you mentioned with your comment section. The podcast world, in a similar way, is very broadcast oriented. I don’t have a great solution for that.
Anyway, you’ve got a lot of amazing stories that people bring on the show. At the same time, I also sense some skepticism in how you’re talking about these things. How are you balancing that? We’re excited about this technology. You’re excited about what it could do. At the same time, there’s a disconnect between what people are saying that their technologies are capable of and what they actually are capable of. How do we reconcile those? Or is it necessary to?
Todd Bishop: I tend to think of this on the individual level, and the organizational level, and the societal level. Where I get into trouble sometimes is just how much of a tech enthusiast I am. This has actually been a fundamental problem for me as a reporter over the years, because I go in and I’m like, “What can your magic do?”
Implicit in that question is “What’s the positive? What’s the good that your magic can do from my perspective?”
I’ve had to learn. You mentioned reader comments, I mean, reader and listener comments. Part of that has been harsh feedback on me being too optimistic about one company’s or another’s potential to do something with [its] technology.
So I’ve learned that over the years, and it’s become a bit of a reflex for me as a journalist. And obviously, it should be there all along, and it is to a certain extent. But a lot of the best journalists are cynical. I’ve had to sort of push myself to be even more cynical in the tech world and get past the flashy new thing that I can’t wait to try.
Sam Ransbotham: Todd, thanks for taking the time to talk today. It’s been fascinating. For listeners who thought this was as interesting as I did, head over to GeekWire and listen to another conversation that Todd and I had. Todd, thanks for joining us.
Todd Bishop: My pleasure, Sam.
Sam Ransbotham: Thanks for joining us. Todd and I had a lot of fun with that conversation, and I hope you enjoyed it as much as we enjoyed recording it. Head over to the GeekWire feed for the companion episode. Next time, I’ll be joined by OpenAI’s first chief economist, Ronnie Chatterji. See you then.
Allison Ryder: Thanks for listening to Me, Myself, and AI. Our show is able to continue, in large part, due to listener support. Your streams and downloads make a big difference. If you have a moment, please consider leaving us an Apple Podcasts review or a rating on Spotify. And share our show with others you think might find it interesting and helpful.