Demystifying AI with Jeff Wilser

00:05

I’m Eric Chemi and this is Politely Pushy. Jeff Wilser is my guest today here on Politely Pushy Jeff thank you so much for joining me for I have to start you are currently the host of the podcast AI-Curious but it seems like you started it before generative AI took over the

00:23

entire world so how are you so so uh so much how did you have so much foresight to see like okay AI media is the place to be first off great to be here uh super excited second off you’re giving me way too much credit I cannot claim to be years and years ahead of the curve like

00:41

everyone else I was or like many folks I was fascinated with the possibilities of generative AI I started playing around with ChatGPT doing things like asking it weird prompts like what if Abraham Lincoln somehow discovered an iPhone that had traveled through time and how would that

01:05

impact his presidency in a civil war and write it in the style of Ernest Hemingway and kind of played around with prompts like that and I was kind of amazed at what it was able to pull off and that’s kind of what got me obsessed with all things AI but your background is not

01:25

necessarily AI right when I was researching this podcast I was expecting okay if he’s got this AI expertise you know maybe he was a developer or AI startup and I’m looking at you know celebrity types of outlets that you’ve worked on as a journalist and and all

01:41

kinds of fun let’s say less technical things so how how did you make the jump to AI. I have zero technical background yeah no developer chops no I cannot I I will flunk any serious conversation that gets deep into the weeds here on the tech front um no great question I I’m a

01:59

general interest writer in general interest journalist I’ve written eight books that have ranged from Firefighters to politics to health trends to business to tech um and I write for magazines ranging from Time Magazine to GQ and so on and over the years I mean I have kind

02:22

of commitment phobia and I lose I switch focus very quickly so I am constantly going from topic to topic to topic and in a way that’s what intrigued me about AI is just how multifaceted the topic is and this is the first time in my 15ish year writing career that I kind

02:48

of forced myself to pause and say wait a second slow down this is a topic worthy of spending serious serious time on in putting a a lot of chips onto the table for I became fascinated by how in a way AI is a general interest topic it’s not just Tech I am uh enthralled by how it

03:13

has an impact and has uh it is involved in topics ranging from art and creativity to obviously the economy and the workforce to politics to geopolitical military issues to business and productivity and on on and on it’s hard to find a topic that or a

03:36

bucket of life that AI is not somehow involved in and that’s one reason why I was so drawn to kind of covering it and following it more closely you know it’s funny you mentioned you’re afraid of commitment and I was gonna ask do you mean commitment in like personal life or

03:51

just that’s a whole other podcast right there Eric but so I wasn’t gonna ask that question but then on your LinkedIn I go way deep and it says you’re the founding editor of The Plunge the leading website for grooms like that’s the ultimate commitment type of story and yet you’re

04:06

telling me you’re afraid of commitment so so what was that like running very different chapter of my life but the idea at the time was we wanted someone who was and let’s face it it was a bit absurd I was single at the time and overseen in publication writing about articles of

04:27

how to be a groom so this is me bachelor in New York writing about wedding planning and wedding etiquette and decorum so have you never been the main I had never been a groom I had the time had no plans to be a groom yet here I was writing about being a groom I was doing things with

04:53

the knots and I was for a while they kind of like groom expert at the knot which was again so you for folks who don’t trust who have the media this is exhibit A right here wait a second why is this guy why is a journalist writing about grooms who’s not a groom so you

05:11

know our our our uh at the time our thinking was well this is good it gave me kind of an outside perspective fresh eyes unbiased right because if you are a groom then you’re shading everything just to kind of like skew towards what you did with your own wedding I was

05:30

saying hey I’m objective I’m just a neutral Observer and I can write with dispassion and objectivity is that the approach that you take to AI now it’s like I wasn’t a groom but I’ll write about Grooms I’m not an AI guy but I’ll be an AI Outsider and try to talk about it from that point

05:48

of view that’s a actually pretty great question and you’re kind of forcing me to be introspective here um in some ways there is I think the common all cutting aside I’m mostly joking around with The Plunge stuff but all kidding aside the one common thread in all of the work I’ve

06:07

done as a journalist um ranging from writing about cryptocurrency to Firefighters to being groom is I try and take a kind of dense or opaque topic in subjects and make it engaging right so on the surface AI and Ice sculptors at your wedding reception

06:34

have nothing in common but in a sense most men to use a stereotype that I think has least a bit of truth to it have not spent a lot of time thinking about what goes into planning a wedding so there was an art of taking a topic that they know nothing about and trying

06:57

to demystify it and I try bring the same approach to like I I’ve written about blockchain Bitcoin and crypto a ton over the past six years that can be a very opaque complicated in the weeds topic and I try and use storytelling and focus on on humans and

07:20

interesting uh compelling juicy stories people can relate to and people find compelling I use that as a way to unpack and get access to trickier topics so I try and take that same kind of curiosity and openness and willingness to really engage on a subject and use that to

07:44

cover any topic whether that is weddings or crypto or AI how much is your focus now only on let’s say Ai and crypto versus all the other stuff that you had done in the past are you still doing general interest or are you all in on technology right now uh no I definitely

08:02

am still doing general interests that and I think I will I’m guessing I will always be doing that I I I do I do have a I’m a I know it sounds cliche but it’s true I’m I’m a curious guy and whatever happens to spark my interest um I might follow that path so I’m not closing off

08:21

other types of stories I still write about travel I still write about exploration um I more most recently finished finished a book about the most uh finished a book about the Explorers Club which is a the kind of distinguished organization founded in

08:42

1904 of the literally greatest explorers in the world ranging from the folks who originally were trying to find the North Pole South Pole then included John Glenn and the Apollo Astronauts and eventually Jane Goodall and leaders in conservation and climate change science so this book

09:03

absolutely captivated me and I it was one of the honors of my career to interview titans in the world of exploration and new research going back from the 1880s all the way to to day for in in NASA scientists so that has very little to do with AI or Tech and I just

09:22

ate it up and continue to work with the Explorers Club so um while my large Focus is on AI and Tech I am absolutely uh still trying to embrace that kind of wider general interest open curiosity exploration mindset as well what is the business model for someone like you

09:42

right so self-employed freelance speaking writing books doing a podcast how do you balance it all and what’s the flywheel look like of how one form of content brings in the next form of content yes yes yes yes and yes I found as a as a self-employed freelancer I

10:00

think most Freelancers can relate to this the more you just do only one thing the riskier that is I’ve been very lucky over the years to have great relationships with editors with book publishers um so for at a high level the way it works for me is to have large

10:23

anchor clients and anchor chunks of work whether that’s book writing or I also do collaboration on books and ghost writing uh having anchor freelance writing contracts I think that for you know I have tons of friends who are Freelancers and I’m sure you do too and when folks

10:46

try and Cobble together a living just from all right I’m gonna write these articles here and those articles there and you Cobble together these checks and even for good let’s let’s say let’s speak plainly here right let’s say you write a like a in the world of

11:06

freelance writing a you know long form piece for a prestige publication that pays something like 4K 6K 8K which in this world is a lot of money for freelance writing it requires cobbling together a lot of those to make the math work to actually make a living so for me

11:28

I’ve always found that it makes sense to have a mix of Articles and then larger projects as well and that that kind of that overall portfolio if you will um makes more sense than doing just one thing yeah I think about that too like let’s say you round number 5K you get

11:45

for one one piece how much work did you put into that piece and how many of those are you getting in a year and and there’s no guarantee you’re getting any future work and how much marketing or self you know self- selling you have to do to continue to get them right that’s

12:01

a tough way to make a living if you’re just picking up that that piece work one by one without I see your point like having those big anchors maybe bigger projects I you’ve even uh run the consensus conference for example right you’re doing a lot of different kinds of

12:15

things it makes me wonder about the future business model for journalists in a world where like coming from my world no one’s watching cable TV anymore right so what’s gonna happen to all of those people right how do they use their brand to build something for them that that is

12:30

sustainable in the modern world yeah I think it’s a valid concern I think that I mean well let me back up for a moment I think that folks are rightly worried about how AI will impact the media landscape and frankly how AI will impact every industry is a valid fair question it’s

12:52

one I’m asking myself I think people lose sight of the fact that even before AI there were grave macro concerns about media to begin with like to be blunt the media Pie has been shrinking long before AI entered the scene right look at the state of local newspapers look at

13:19

magazines right we we we don’t look most most websites we don’t have the same media rich soil now we did 20 probably even 10 years ago and so to your point Eric I think even before AI it is the the uphill slope that journalists need to are are are running up you could

13:46

argue that is becoming steeper and steeper and steeper and likely with it’s open question of how AI will impact this right because potentially that gets steeper still maybe for those who are leaning into it AI can be a bit of a tailwind to help you in that battle to

14:04

mix metaphors but um yeah it’s a it’s a tough question I certainly don’t have any any solid answers yet I’m not sure anyone does how much AI are you using in your own work like are you using an AI podcast editor AI for writing AI for maybe marketing how much AI tools do you

14:20

personally use yeah not a day goes by that I don’t use AI in at least some capacity um that’s often ChatGPT but not always um just to give you a couple of examples I I certainly for the podcast I use you the editing software I use is has AI embedded throughout um

14:42

shout out to the Scripts which I know is becoming more and more popular Scripts is the one I keep hearing about that’s the one that’s become that’s is a reason why it is the buzziest uh podcast editing system right now they have gotten whole hog embraced AI and many other

14:60

elements and I use that liberally not not always there are some things that if you’re not careful like first for not to get into the weeds here but one example I think is illustrative they have a feature that lets you um press a button and all of the like gaps of silence in a podcast

15:21

can be cut out or they have a feature where all of the ums likes younos are instantly removed and and you can kind of go through more granularly say I want to take out the ums but not the likes or vice versa if you’re not careful if you paint to Broad of a brush and just say

15:43

yes strip them all out it goes a little overkill goes haywire and sounds unnatural right and there’s also something to to humans like you and I speaking when I make little slip UPS which is quite common in my case there is an element of authenticity there right and when things seem a

16:06

little too smooth and polished it feels less real and that undercuts what many folks like about podcast to begin with right so I think I know we’re getting a little nerdy on this one example but I think there’s a larger takeaway about AI and that if we’re not careful in overuse of

16:29

it can kind of strip out some of the human authentic imperfections and can seem a little too kind of shiny and plasticky it makes me think about when you were giving the examples here of you saying um and like I was thinking about if you ran that through

16:47

the filter would it blank out you saying those words right your description of what descript does would eliminate your saying of it removes the ums in the legs would it just say it removes yeah or it wouldn’t be silent it it keeps it would kind of compress it all

17:05

say it moves this and this and this and this I’m like what the hell did he just say right so and I’ve done things like that where I I’ve I’ve I backed off it but for an episode or two I was experimenting with this tech and I overdid it and like the guess and I I

17:22

relisten to it and there were important moments important that the woman was was saying and I think it it over it was overeager and uh I forget the exact example but it it took away some crucial context I went back and I fixed it thank goodness but yes that’s just one example of how AI

17:41

can go kind of awry um but to answer your original question of what else I’m using it for um I do not use it to write for me anything of imports um part of that perhaps is professional pride and stubbornness right I’m a I’m a paid writer so I think no I can do this

18:04

myself I don’t I don’t want you to write for me I can write better than you which I think is still true for the moment that might change um but what I do think I use it for liberally is all kinds of brainstorming as a thought partner I use it for pre-research or to help with

18:23

research um given its fondness for hallucinating of course I don’t use that as the final research but let’s say I am writing I’m gonna make something up if I’m writing an article about the best museums in Colombia maybe I’d ask it to tell me all about museums in Colombia that is

18:49

something I could then use and I would check everything I’d vet it find other sources but it’s nice to have a little bit of an easy primer right um I also think that it can be a great um uh it’s great for helping you find blind spots and weaknesses so if I have an outline I

19:09

might send it to ChatGPT ask for its thoughts hey what am I missing what are what are things I’m not thinking about if you had to double the length of this what would you add um push back on me right so I think it’s very useful to give you perspectives you

19:28

haven’t thought of um I on my podcast uh AI curious I interviewed and i’ I’ve talked to multiple CEOs and thought leaders who do a similar version of this so I’m just stealing it from them um one person I spoke with who is entrepreneur if they’re writing a pitch

19:47

deck they would say hey put yourself in the role of a skeptical investor let’s say you are a VC who’s really a hardnose and asking tough questions what questions would you ask me what push back do you have so you can use ChatGPT to anticipate uh anticipates kind of the other side

20:09

and can help strengthen your arguments in advance and see I’m slow there I’m slow getting back to my question and moving my window around that where descript would have would have cleaned up that little that little good this is real this is a free flowing real authentic

20:25

conversation this is proof of humanity here I was writing notes down that’s why I was was writing one more thing I’ll buy you some time then one more thing I use it for a lot I I now and this is not this I’m not like this is not uh breaking news crazy unlock idea no one

20:40

thought of before but just what I do in my my day-to-day it is now I usually start a search with AI as opposed to uh Googling right so and it’s It generally gives me a result in a cleaner easier to follow format than just um bringing up like a seven Google seven sponsored

21:07

results right exactly it makes you wonder about that business model though right that if we’re skipping to the ads can ads in infiltrate into the AI answer like what what’s gonna happen they’re gonna have to no I mean it is a existential question for the internet

21:21

you’re spot on um that no no one really has an answer here in a sense that like the entire Revenue model of the internet is built upon search and ads and when AI powered search blows that up when just be to be very concrete here if um back to my stupid I don’t

21:44

know why my mind went here back to my example of the museums in Columbia is that an article that you have worked on or will be working on no not I actually intentionally chose something thing I would like a 0.0 yeah yeah yeah I my mind was trying to find something that

22:01

was nothing I would do so I’m not you know um uh spilling state secrets here so back to my museums in Columbia article that I have not written and will not write in the past I write this and let’s say it’s uh museum.com you know that that’s publishing My article and

22:22

they are hoping that people when they search best museums in Columbia I will that their article written by Jeff Wilser museums on museums.com they’re hoping that will be in a top three or four of results people click on that link they see the ads the ads are for Airlines or

22:41

cruises or hotels or whatever what have you right in the future um instead of people clicking that link when they Google or when they use Perplexity AI search or ChatGPT and they say hey what are the best museums in Columbia able to say here here the top 10 and there’s no

23:02

need to click on my article right um and that is a again I don’t think it’s overwrought or hyperbole to use the sprays existential question for the internet and as an aside Google is in a very interesting place here because they’re kind of on both sides of this

23:22

Google of course is developing their own AI meant to compete with OpenAI but their entire company is kind of predicated on ad revenue from search so are they want you to waste time going through one by one all these results yeah yeah so it’ll be fascinating to see how this shakes

23:44

out so when you were giving that previous answer I was wondering about four questions that come up number one when you started the podcast audience who is the idea what was your idea for who is the audience if I build it will they come was there a demand for this or

24:00

did you just thought hey I’m going to just start this and see what happens but who are you trying to reach yeah great question I don’t know that I and perhaps I should have I don’t think I was quite as uh explicit like in a conference room with a whiteboard kind of carving up

24:19

demographics and talking about who you know what age and this and that I was following my gut of thinking I am personally very curious about what AI is beginning to mean to the world and how it will evolve and where this will all go across a multitude of dimensions right ranging

24:44

from business to philosophy like I’ve had on like neuroscientists to talk about how will we know if AI becomes sensient like what are the tests right that’s one episode I’ve had then I had an episode with Hollywood producers on how AI is impacting the movie industry

25:07

in ways good and bad challenging promising terrifying exciting all that stuff right then an episode with a lawyer on the legal impacts point is in getting back to your question I had confidence in my own curiosity of all this so my leap of faith is that others

25:29

would too there must be others like me that frankly don’t want to go crazy deep in the tech right I’m not gonna have like a two-hour episode getting into like inference and the the kind of like the the Deep machinations what’s that sorry I was gonna say like the examples of how

25:51

you’re tweaking yeah your inference model and and how many layers are in your neural network and what’s your waiting and zero and like like how crazy do we want to get with this stuff right exactly maybe someday but for for now we’ll leave the waiting questions to

26:07

different podcasts and so my target audience really like folks like me that recognize it as important but don’t want to spend too know and obviously I am spending more time on it you know I need to to develop the content of course I am personally now spending a lot more time

26:23

in AI lands but I view my role as helping curate interesting stories and themes and topics that I think that someone who has a general interest interest person like me they they’re they’re reading the news right these are like New York Times readers they are savvy you know staying

26:43

up on current news current affairs they get the AI’s importance they want to know a bit more without like devoting crazy time to it my I try good so I I thought you were done I need the script to tell me when when the pauses are I’m just going to say do you Outsource any of the

27:03

editing or the booking or the production or are you doing everything here as of now I am doing it all I obvious that that will you know hopefully/likely change in the future but I and part of actually just to peel back the curtain a little bit here um which I think is a

27:20

big part of your of your podcast is great it’s fun it’s fun to and I’m guessing Eric you can relate to this as learning new skills I think that for journalists a lot of the time we’re doing the same thing again and again and again and writers are doing the same

27:39

thing now I’ve there’s always new things you’re writing about there’s always new topics and personalities and that keeps things fresh invigorating sure but it’s rare you’re learning like a new skill and there’s something just purely pleasurable about learning things it’s

27:58

like oh I need to learn audio editing I need to figure out the the basic mechanics of this which platform to use what are the shortcuts that’s it’s it’s it’s especially I think a larger theme of mine in life is I think it’s rare that people pick up new skills and interests as they get

28:24

older one pet theory of mine is we tend to like peak of our interests breadth let’s say in college and after that our what begins is a very broad range of interest in things we do things we are investigating and learning and practicing whether that’s a musical

28:47

instruments or picking up a new sport like tennis or photography whatever it is we start Broad and as our career develops as we get more focused on relationships or family those interests tend to narrow and hopefully deepen but narrow and so when I can and I I wish I

29:09

was better at this than I am but I like taking the opportunity to go broad again to to do new things things that I don’t need to do but that enriches my life in surprising ways when you’re booking the what’s the process for how you find guests and what makes for a good guest

29:26

on your show great question there there’s I found that the best episodes and guests I’m not going to name names in either direction are the ones that and going back to the thesis of the podcast AI-Curious I am either truly insanely curious about this particular person or

29:55

this topic so I’ll do one of two things there times when I see some AI notable person whether on Twitter/X or what somehow in the news I’m like okay I need to talk to them so I’ll give one example uh of this uh There Was You probably saw it a month or two ago these two Harvard

30:17

kids who Harvard students who created this wacky and wilds and frankly deeply disturbing tech device by they hack together Rayban smart glasses that kind of have like a camera on it and then used a reverse facial recognition software and then used AI to

30:41

allow them in near real time not quite but near real time as they’re walking down the streets almost like literally Black Mirror style they can like walk they they would see they saw a guy in the subway and within a minute they realize okay that is Joe Schmocetelli who

30:59

listes this address who does this and they could tell Joe Schmocetelli um oh aren’t you Joe Schmocetelli who lives at you know 214 uh Main Street and it was wild and that story went went viral it was on the front page New York Times and so when that was forwarded to me by some friends

31:22

I was like okay I need to talk to these guys and so I invited them on the podcast and also I want to say as a as a PSA for these uh these two Harvard students and they’ve kind of gotten a bad rap and as they as the as one describes on my podcast they did not do

31:37

this to create some dystopian monster he knew full well this raised all kinds of privacy concerns and he in in fact the whole purpose of the experiment was to raise those questions they wanted to get this out there so humanity realized holy this kind of chilling technology is

31:57

not something that will scare us in the future it’s already here today so we should figure out guardrails um anyways a long windway of saying sometimes I will find the person and like okay I want them on the show that’d be a great conversation other

32:13

times I’ll back into that through the uh a topic like oh I really want to do an episode on AI and the movies who is a good guest for that and so I’ll kind of back into it by focusing a topic But ultimately in either direction I am generally following my own curiosity

32:38

whether that’s topic or person I see I remember that story that was crazy I I read a lot about that uh the facial recognition what you could do with it because I do think they’re just showing that it’s possible right hey if two college kids can whip this up so quickly

32:54

imagine what’s possible when giant governments giant trillion dollar organizations want to put an army of tech workers behind this what else could they do Eric it took them four days four days to do this two college kids four days like what what is possible then

33:09

with here there’s a billion dollar budget go do what you want to do I don’t think it’s weird as much hype as there has been around AI and there’s been a ton and in a way I’m I’m guilty of feeding into it I have I’ve had I’ve fed the hype beast myself as much hype as there is I still

33:30

weirdly at the same time think there’s almost an under appreciation of how much the worlds can and likely will change with new AI advancements and that’s one reason why I want to do this podcast is to kind of get ahead of it I don’t have any illusions that I am really part of this

33:52

solution I don’t I’m not going to pretend to have these like altruistic motives oh I’m gonna help us like really figure out how to you know handle AI I don’t have any illusions there um but I do think it’s important the very least to be aware of what’s coming down the

34:09

pike to help us if nothing else emotionally brace for it and kind of seal ourselves for it when do you think you’ll run out of steam on it when do you think you’ll move on as you’ve moved on with your other interest and you’ve sort of got to this point okay I know

34:23

enough about this I’m I’m curious about the next thing how long do you think AI curious has legs how many more years before you think okay I’m good I’m gonna do something else yeah great question honestly I feel like I am just getting started I that question hasn’t even

34:39

you’re the first person to ask that like including myself I haven’t even I’ve been a just past the year mark recently and I have been religiously cranking out episodes every Thursday a new long form interview I pump out I think the one I think last last Thanksgiving I think I

34:60

released that on Friday not Thursday that was my one exception other than that it is I’ve just been cranking this out I love doing it there has not been a moment I’ve thought about oh is it time to I actually I have I I’ll just tease by saying I have pretty exciting plans

35:17

on leaning further into it on scaling up on doing more uh so if there’s ever a moment where I become um restless the answer in my mind at least right now going forward is to be doing more not less I’m I’m I am truly enthralled by this topic in ways I

35:42

have not been in Prior uh projects of mine in my journalism career so you like this better than talking about nightclubs or grooms and weddings you feel like this is a little bit actually I do I want I want to I want to clarify one thing for a moment moments um I this

35:60

is not to say there are no other projects I have not loved as much and to be um I many of the books I’ve written I’m very proud of and I love doing so it’s not to say I like a more than B this more than that but as far as what makes what has legs to continue to

36:23

invest my time and energy in both from a intellectual Pursuit also business model this makes sense in ways other topics have not that I’ve also enjoyed like for instance um working with the Explorers Club is an absolute Peak honor of my career I and I

36:45

continue to work with them I think I’m allowed to say there’s a second book that I am working on with them I am extremely fired up for that and am very excited about continuing that literally that exploration so there are other projects that I love and will keep doing

37:10

but this the unique aspect of AI of how it touches on so many different kind of parts of the world and then also I think by jumping into it on a relatively early uh time frame gives me a position to continue to to grow and and and tell stories and help others connect with AI

37:31

so I will definitely keep leaning into it for the foreseeable future I love it Jee this has been amazing thank you so much for sharing some of your insights today and then the process for how you’re doing all the things that you’re doing it’s amazing to see so we’re gonna have

37:44

to look for that look for that whatever is coming whatever is coming with the the new stuff as you continue to grow and build it out thanks so much Jeff thanks so much I really enjoyed [Music] thank you to my guest and thanks for listening subscribe to get the latest

38:01

episodes each week and we’ll see you next time.