A Cry for Help in Veterinary Pathology, Burnout, Economics, Digitization, and AI with Dr. Eric Fish Dr. Eric Fish, DVM, DACVP, joins Dr. Andy Roark to tackle a tough question: Why does being a veterinary clinical pathologist feel darker right now, and what happens if AI turns “nice to have” diagnostics into “don’t need to have”? They unpack burnout and pay compression (including pathologists earning less than new grad GPs), consolidation in the lab market, and how digitization has pushed cytology toward point-of-care and high-volume “turn and burn” workflows with per-case bonuses that can be as low as $8 to $12. Eric explains how consultative lab medicine has shifted toward internists, why cytology’s cost and turnaround time are eroding its value, and what a better future could look like with AI that augments, triages, and keeps a human in the loop. Gang, let’s get into this episode. Keywords: veterinary pathology, clinical pathologist burnout, digital cytology, AI in veterinary medicine. Links: Substack newsletter All Science Great and Small; LinkedIn. Dr. Eric Fish is a veterinary clinical pathologist and laboratory director at a large private referral center, with a career spanning academia, entrepreneurship, and industry. He co-founded Lacuna Diagnostics, the first point-of-care digital cytology company, spent years at IDEXX as a diagnostician and technology scouting manager, and even found time to work occasional relief ER shifts. Eric brings a unique perspective to the conversation around artificial intelligence: he has not only used voice recognition and large language model AI tools in his own practice, but also worked on teams building computer vision applications for medical imaging. He writes about AI and other hot topics in veterinary medicine in his Substack newsletter, All Science Great & Small.

Welcome everybody it to the corner of Shame Veterinary Podcast.

2

I am your host, Dr.

3

And guys, I got a wild episode today.

4

a cry for help.

5

From your pathology friends, this is a

really interesting conversation with Dr.

6

Eric Fish.

7

He's a clinical pathologist.

8

He writes a lot about pathology, about

science or medicine and about ai.

9

And, he's been dark recently.

10

He's been, he's been.

11

He's been sharing, some views of the

veterinary pathology profession that

12

are, that, dare I say pessimistic.

13

It, it's, it looks like a dark

time for our pathology friends.

14

And I, I just kept looking at

it and going, is that what it's

15

like to be a clinical pathologist

in fact medicine, really?

16

it's a fascinating conversation today.

17

We talk through like, what

are the economic drivers.

18

pathologists and the pathology

work in our profession.

19

How is, veterinary pathology changing?

20

What are pay structures

for pathologists today?

21

And if you're just interested

in understanding how this

22

piece of our profession works,

this is fascinating episode.

23

It's really interesting.

24

And also there's a lot of the

components that Eric talks about.

25

That I could kind of extrapolate

into general practice as well.

26

Pathology is interesting because

it's sort of similar to radiology

27

and that it can be digitized and

has been digitized in a lot of ways.

28

And also the experience the pathologists

have and the radiologists have,

29

they, they're pretty different.

30

And we get into why that is,

is Anyway, guys, this is a

31

really interesting episode.

32

I have not had a conversation

like this, that I can remember.

33

I hope you guys will enjoy it.

34

Let's get into it.

35

Kelsey Beth Carpenter: This is your show.

36

We're glad you're here.

37

We want to help you in

your veterinary career.

38

Welcome to the Cone of Shame with Dr.

39

Andy Roark.

40

dr--andy-roark-_1_05-05-2026_141150:

Welcome to the podcast, Dr.

41

Eric Fish.

42

How are you, my friend?

43

eric_1_05-05-2026_141149: Great.

44

Thank you.

45

I was very excited to be here.

46

It's kinda like, when you listen to

the radio and someone says, "Longtime

47

listener, first-time caller,"

48

dr--andy-roark-_1_05-05-2026_141150:

Excellent.

49

Well, thanks, thanks a lot

for making time for this.

50

For those who don't know you, you

are a veterinary clinical pathologist

51

and a laboratory director at a

large private referral center.

52

you've done a ton of

things in your career.

53

You've been academia,

you've been in industry.

54

You have kind of worked all over

as a clinical pathologist, that's

55

kind of where you come from.

56

you write on Substack, and that's actually

where I became sort of aware of you.

57

It's called All Science Great and Small.

58

And

59

eric_1_05-05-2026_141149: right.

60

dr--andy-roark-_1_05-05-2026_141150:

of different things.

61

yeah, you write about a lot of different

things, just generally in, in veterinary

62

medicine and science in general.

63

You write a lot about pathology

and the profession of veterinary

64

pathology, and that's kinda what

I wanted to talk about today.

65

I know that I'm talking to you, kind

of not as a representative any- of

66

any company, anything like that.

67

you know, this is, this is just 100%

you and me, just you and me talking.

68

So,

69

eric_1_05-05-2026_141149: Absolutely.

70

dr--andy-roark-_1_05-05-2026_141150: yeah.

71

So, l- let's, let's kinda,

let's kinda get into this.

72

I wanna sort of give a, a, a

high-level idea of kinda what

73

I've found most interesting.

74

So you've been writing a lot recently,

and the thing that I've been sort of most

75

surprised by, my feeling is that you are

a bit pessimistic about the future of the

76

career path of veterinary pathologists.

77

And so you, you put a number of

things out about how pathology as

78

a, as a vocation i- is changing,

and specifically sort of how it's

79

changing with AI and things like that.

80

And I gotta tell you, Eric, I was really

blown away as I'm sort of reading through

81

your sort of descriptions of, of how

technology is being used and how it is

82

changing what has traditionally been

done by pathology and, and pathologists.

83

and you kind of paint it, paint it

into sort of a dark picture here.

84

Let me, let me hand this over to you.

85

Can you talk to me at a high level

about sort of the state of clinical

86

pathology as a practitioner?

87

eric_1_05-05-2026_141149: Sure, yeah.

88

So I think what you're referring

to is a month or two ago I wrote

89

something called A Cry for Help, and it

90

dr--andy-roark-_1_05-05-2026_141150: Yes.

91

Yeah.

92

eric_1_05-05-2026_141149: public, our, our

specialties journal, Veterinary Clinical

93

Pathology, there was a large survey done

of European clinical pathologists, so

94

dr--andy-roark-_1_05-05-2026_141150: Okay.

95

eric_1_05-05-2026_141149: to the US,

but a lot of themes seem very similar.

96

In, in very high levels, there was very

high levels of burnout, dissatisfaction.

97

There were a lot of complaints about pay.

98

And when you dig it down into the

detail, a lot of clinical pathologists,

99

particularly in Europe, but also

frankly in, in North America, they're

100

starting to make less than even

new grad general practitioners.

101

And, you know, GP's work very hard.

102

They generate a lot of revenue,

so that's nothing against them.

103

But

104

dr--andy-roark-_1_05-05-2026_141150: Sure.

105

eric_1_05-05-2026_141149: cost spending

four to six years in advanced training,

106

and then you're making less than someone

who got their DVM six months ago.

107

So it's kinda crazy.

108

And there's a number of reasons for this.

109

It's not one.

110

It's very multifactorial.

111

part of it is clearly a story about

consolidation, particularly in the

112

North-- in North America, there's a

handful of big labs that are probably

113

eighty-five, ninety percent of the market.

114

And then there's a few boutique labs,

a few people work in universities,

115

but the, the main point of

action is the private lab sector.

116

So anytime there's consolidation,

they have more power to set wages

117

and policies, things like that.

118

Part of it's definitely

a story about technology.

119

Some of that is the transition to

digital, which I was a part of.

120

I was involved with some companies

that did digital cytology.

121

That has had a lot of positive impacts,

but there have also been some growing

122

pains and disruption to the industry.

123

So I think that's part of it.

124

And AI is starting to creep in.

125

I don't think it's at the point yet

where that is the main reason for some of

126

dr--andy-roark-_1_05-05-2026_141150: Okay.

127

eric_1_05-05-2026_141149: on the horizon.

128

And there's d-- there's at least a couple

companies out there that are starting to

129

market point-of-care analyzers for dig--

for cytology using AI to interpret them

130

really without a person in the loop.

131

And those are starting to get traction,

and I think that as that grows and

132

they add more capabilities, it is gonna

start to eat into the lab business and

133

likewise, the employment of pathologists.

134

dr--andy-roark-_1_05-05-2026_141150:

Talk to me a little bit about

135

how the day-to-day work that you

do as a pathologist has changed.

136

And so when did you, when

did you become boarded?

137

eric_1_05-05-2026_141149: did my residency

from twenty thirteen to twenty sixteen,

138

so got boarded in twenty sixteen.

139

I did stay on and do a PhD, so a

lot of path residencies are, are

140

actually combined with graduate

training, particularly PhDs.

141

So I finished that in twenty

nineteen, but I got board certified,

142

by the ACVP in twenty sixteen.

143

dr--andy-roark-_1_05-05-2026_141150: Okay.

144

But you were actually working

towards being boarded as a

145

pathologist in like 2013.

146

Just in your career, so you're

th- 13-- over the last 13 years,

147

Eric, tell, what has changed for

you in terms of doing this work?

148

When you started your training, how

different is it now from the type of work

149

that you were either doing or anticipated

doing when you started your training?

150

eric_1_05-05-2026_141149: Sure.

151

So there's a couple

different dimensions to that.

152

You know, starting training in a

university setting, I was at Auburn,

153

and I stayed on for faculty there.

154

It's a much more diverse role

when you're in a university.

155

Part of what you're doing is

reading cytology and hematology

156

slides, but you're also teaching.

157

You may be doing research.

158

You may be consulting with the

technologists who are running the

159

chemistry panels, things like that.

160

You may be having conversations

with clinicians for consults

161

on weird test results.

162

Whereas in the labs, your job is pretty

much ninety-seven percent just reading out

163

cytology slides, almost like a machine.

164

And that was one of the complaints

they called out in that European

165

wellness survey, is that they did not

feel like they were utilizing their

166

breadth of training and experience.

167

So that's partly a sector

thing versus new technology.

168

the biggest impact I've seen

really since starting, setting

169

aside AI, is the digitization.

170

So that has pushed more and more

cytology to the point of care.

171

Instead of mailing things into a

big lab network, you may be having

172

a scanner in your hospital that

beams it up to a, a contract group

173

of pathologists in the cloud.

174

And that can allow faster turnaround

times, but one of the things

175

is not all the scanners can get

anywhere close to 100X resolution.

176

So there are some optical limitations.

177

So unlike digitization for radiologists,

where the transition was pretty smooth and

178

there weren't really any handicaps from

the technology, in cytology, there are

179

some things you just can't see on digital.

180

So that has been a little

bit of a rocky transition.

181

would say another big change has been,

frankly, the economics and the price.

182

We're starting to see, you know, even when

I take my own pets in and they recommend

183

cytology, it can be two hundred and fifty,

three hundred dollars to the end client.

184

dr--andy-roark-_1_05-05-2026_141150: Okay.

185

eric_1_05-05-2026_141149: some of the

things that happen with cytology is it's

186

best used as a quick screening tool.

187

So is it this or is it that?

188

But when it takes seven days, ten

days to get your result back and it

189

costs as much or more than a biopsy,

you're really starting to hurt the

190

value of that rapid screening test.

191

So I think some of the volume issues

and pay issues we're seeing with

192

clinical pathology are really due

to those macroeconomic changes.

193

So it's like the cost has gone way up,

and has been harder and harder to get

194

some of those things in the turnaround

time where they're most useful.

195

So I think some people are just

opting out of sending cytology, so

196

even before you get to the AI piece.

197

dr--andy-roark-_1_05-05-2026_141150:

Really?

198

Okay.

199

So there's, there's a lot of

things here to sort of unpack.

200

eric_1_05-05-2026_141149: Sure.

201

dr--andy-roark-_1_05-05-2026_141150:

Talk to me a little bit more about

202

the transition of pathology towards

this focus on cytology slides.

203

You know, and so this is kind of something

that, that you wrote a bit about of,

204

you know, 97% of the, of the work now

is just looking at cytology slides.

205

I still think of, you know, histopath

and s- you know, biopsies and

206

things like that, but it, it sounds

like that has sort of changed.

207

Talk to me a little bit about that.

208

My impression is just s- sort of from,

from reading through some of the patie-

209

papers, there's some frustration around

the, the, the idea that basically we

210

just, we're the readers of slides now.

211

Talk, talk to me, u- unpack that for me.

212

eric_1_05-05-2026_141149:

Yeah, absolutely.

213

in a perfect world, clinical

pathologists are supposed to be

214

experts in laboratory medicine.

215

So we're supposed to know how each one

of the parameters on a chem panel works,

216

when they go wrong, what's causing the

error, talking about interpretation.

217

A lot of that has sort of, in

an almost turf war way, been

218

absorbed by other specialties.

219

So internal medicine is often the go-to,

and many of these lab companies actually

220

hire internists to be doing some of the

that I think in a previous era would have

221

been handled by clinical pathologists.

222

So

223

dr--andy-roark-_1_05-05-2026_141150: Okay.

224

eric_1_05-05-2026_141149: of some scope

creep and turf creep between those there.

225

and yeah, I think that's the major part

for, like, why we're not doing a lot of

226

that consultation on the lab diagnostics.

227

And that's unfortunate 'cause it's

like a third of the training and

228

the content that we go through.

229

It's a third of our boards,

and you just never use it again

230

once you finish residency for

most people who leave academia.

231

dr--andy-roark-_1_05-05-2026_141150:

Why is that?

232

Like, w- so yeah, why, why, I mean,

if, if, if clinical pathologists are

233

trained for it and there's, you know,

there's people around to do the work,

234

like why, why has that-- why have they

been kind of moved out of that role?

235

eric_1_05-05-2026_141149: Yeah,

I think there's just a perception

236

that there's not the value add,

which is unfortunate, 'cause I think

237

there's a huge amount of value add.

238

I think clinicians are like, "Well,

I can interpret my own lab work," and

239

internists are there for the kind of the

higher level consultat- consultations.

240

And it just-- it's, it's actually

pretty similar to cytology.

241

A lot of people have the

perception, "Well, I can do

242

cytology in my clinic for $10.

243

Why am I mailing it to a lab for this huge

price that the client doesn't wanna pay?"

244

With histopathology and anatomic

pathologists, no one has a microtome,

245

no one remembers histology.

246

It's all kind of a nightmare from vet

school, so you're like, "Of course

247

you have to send it to a specialist."

248

The norms and the mores for

radiology have shifted very heavily

249

towards, "You should probably have

a radiologist sign off on this."

250

it's obviously not available everywhere,

and sometimes clients don't consent,

251

but the norm has really shifted there.

252

Whereas I think clinical pathologists

are viewed right now as a nice to have,

253

and I think we're sliding towards maybe

a don't even need to have in the future,

254

dr--andy-roark-_1_05-05-2026_141150: Wha-

255

eric_1_05-05-2026_141149: have

been a must-have in the past.

256

dr--andy-roark-_1_05-05-2026_141150:

why do you think people perceive

257

radiology so differently?

258

Like, why isn't it, those are

the same thing of, you know…

259

it-- in both of these cases, it's kinda

like, well, we've got some information

260

and I'm making a call based on this.

261

You know, having someone who can,

who can verify or validate, there,

262

there seems to be value there.

263

Like, why, why don't

you perceive the, the…

264

Why, why does, why is radiology

going in, in a different

265

direction in people's minds?

266

eric_1_05-05-2026_141149: Yeah,

I'm not really sure about that.

267

I think their, their journey from the

traditional dipping films and chemical

268

vats to digitization has played out a

little differently than pathology for

269

a variety of granular detailed reasons.

270

I think part of it is there might be

more of a shorter loop on feedback.

271

So you take an X-ray of something,

272

dr--andy-roark-_1_05-05-2026_141150: Yeah.

273

eric_1_05-05-2026_141149:

surgical or there's nodules, and

274

pretty soon you're often gonna

275

dr--andy-roark-_1_05-05-2026_141150:

That's right

276

eric_1_05-05-2026_141149: there's

just-- that, that would be my

277

guess is that there's not as

much of a tight correlation.

278

Like, you know, r-radiology

plays really nicely with surgery.

279

So a lot of those things, you're

gonna go to surgery and you saw

280

it or you didn't, and then you're

gonna refine your own mental model.

281

When I saw this in a patient, it was that,

and then it was confirmed when I took it

282

to surgery or referred it for surgery.

283

So that would be my guess.

284

I, I don't know.

285

I, I do think partly for cost reasons,

a lot of people don't consent to

286

histopathology, so a lot of times whoever

looked at the cyto is the last word.

287

And if that was someone who looked

in-house, they think that they were right.

288

dr--andy-roark-_1_05-05-2026_141150: Yeah.

289

Yeah.

290

It's like, let's just call, let's

call it right and no one will

291

ever come along behind us and, and

correct us, so we're taking the win.

292

Okay.

293

when you look into the future,

where do you see this going?

294

Do you think that there's a possibility

that, you know, like maybe, maybe

295

shortened turnaround time on pathology

is gonna kind of thrust it back into, i-

296

into a, a more, you know, favorable light

in terms of, you know, what we, what we,

297

how we use it in, in cl- in the clinics?

298

Or do you think that AI is

going to further kind of push,

299

push pathology down the list?

300

Well, it, you've, you've kind of, you've--

I've seen you sort of talk both ways.

301

I mean, I think you clearly really

enjoy being a pathologist, and also

302

you seem to have some, some concerns.

303

Where, where do you

really think this goes?

304

eric_1_05-05-2026_141149: Yeah.

305

So I'll, I'll-- a couple years ago, I

did a talk on AI and medical imaging,

306

and it was focused on pathology, but

also talked a little about radiology.

307

And I had a slide in there

that said, "I am ambivalent

308

about artificial intelligence."

309

And I think a lot of people when they hear

the word ambivalent think someone means,

310

like, disinterested or they don't care,

when really if, if you break it down, ambi

311

is both and valent is a strong opinion.

312

So it means you have conflicting

and equally strong opinions.

313

dr--andy-roark-_1_05-05-2026_141150: Yeah.

314

eric_1_05-05-2026_141149: of positive

and negative thoughts about AI.

315

you know, if you're a fan of "The

X-Files," one of the things Mulder

316

used to say is, "I want to believe."

317

I wanna

318

dr--andy-roark-_1_05-05-2026_141150:

I want to believe.

319

eric_1_05-05-2026_141149:

makes us more efficient,

320

dr--andy-roark-_1_05-05-2026_141150: Yeah.

321

eric_1_05-05-2026_141149: patients by

getting better at catching things that

322

you wouldn't have seen on the X-ray.

323

It makes us enjoy our jobs more because

it gets rid of some of the scut work,

324

like just kinda doing mindless counting

or mitotic counts, things like that,

325

and really helps us focus on the thought

work and providing care for patients.

326

What I am seeing over the last few years

Is not a lot of data to support this.

327

I'm seeing what little data emerges

that's independent, the performance is

328

often a little lackluster in my opinion.

329

And the model we're moving towards is

not augmentation, it's replacement.

330

So we're pitching boxes that you

can buy from-- for $5,000, or you

331

get it into your lab contract.

332

And for a low consumable cost,

you can get a pathology read.

333

And that is going to, I think, be very

attractive to a lot of practices in

334

this environment where a lot of clients

are very concerned about costs, people

335

are concerned about turnaround time.

336

So don't think that future is inevitable,

but I'm worried that's the model a lot

337

of the big stakeholders are looking for.

338

And, and frankly, it's because

they wanna reduce their own costs.

339

dr--andy-roark-_1_05-05-2026_141150:

Oh, sure.

340

eric_1_05-05-2026_141149: a pretty big

cost center for a lot of these major labs.

341

So, you know, you don't wanna infer

negative, motivations on the part of some

342

of these people, but I, I do worry that a

big part of it is about cost savings more

343

than medical quality or efficiency per se.

344

dr--andy-roark-_1_05-05-2026_141150: Yeah.

345

there's-- the thing in my mind is, you

know, you're not gonna lose your job to

346

AI, you'll lose your job to someone who's

using AI, and it's like, I imagine…

347

eric_1_05-05-2026_141149: people say that

all the time, and it's just become such a

348

dr--andy-roark-_1_05-05-2026_141150:

it's so

349

eric_1_05-05-2026_141149:

kind of annoys me

350

dr--andy-roark-_1_05-05-2026_141150:

It's-- Oh, it's so-- I think

351

it's, I think it's super annoying.

352

It sa- it means nothing.

353

Like, there's no real wisdom there,

and people say it all the time.

354

eric_1_05-05-2026_141149: and

it's really just kind of a frankly

355

self-serving kind of, aphorism.

356

You're like, "Well, Skynet's coming, so

I might as well help it get here faster."

357

dr--andy-roark-_1_05-05-2026_141150:

Yeah, totally.

358

eric_1_05-05-2026_141149:

I don't know about

359

dr--andy-roark-_1_05-05-2026_141150: Yeah.

360

Oh, it, it's definitely, it definitely

write, writes off sort of your,

361

your role in the whole play of like,

362

eric_1_05-05-2026_141149: Yeah, you

363

dr--andy-roark-_1_05-05-2026_141150:

this is…"

364

eric_1_05-05-2026_141149: hands.

365

dr--andy-roark-_1_05-05-2026_141150:

It, it, it really is.

366

do you think that, efficiency in

pathology has been increasing?

367

Do, you know, do you feel like…

368

eric_1_05-05-2026_141149: Absolutely.

369

I mean, if you look at the data,

I-- some of my journey, I've worked

370

at a couple corporate labs, and the

internal turnaround data, obviously

371

can't share, but it's quite fast.

372

It's minutes per slide.

373

I mean, the idea-- the, the rate

that we train in a residency for

374

your board exam when the format was a

little different and you had to read

375

a whole flat of slides was about 12

minutes a case, and you had one slide.

376

dr--andy-roark-_1_05-05-2026_141150: Okay.

377

eric_1_05-05-2026_141149: a

pretty challenging case, but

378

you had 12 minutes per slide.

379

That was kind of the benchmark rate.

380

It's well below seven to eight

minutes in most of the labs.

381

And a lot of them are you've got

pathologists reading four or five minutes.

382

It is lightning fast.

383

And that's part of the burnout

piece, is people feel like

384

they're just turning and burning.

385

They're not putting the thought

386

dr--andy-roark-_1_05-05-2026_141150: Yeah.

387

eric_1_05-05-2026_141149: You're

like, "Well, maybe I missed

388

something on that case, but I do

not have time to call the client.

389

I don't have time to look something up

and give it a little bit more thought."

390

and I think for people in a certain

part of the profession, they

391

may turn towards boutique labs.

392

You're like, "I wanna work

with a pathologist where I

393

have a personal relationship."

394

Maybe the volumes are lower,

the price is higher, but that

395

doesn't work for everyone.

396

So most people are kinda pigeonholed

into one of the big labs,

397

dr--andy-roark-_1_05-05-2026_141150: Yeah.

398

eric_1_05-05-2026_141149: very high volume

399

dr--andy-roark-_1_05-05-2026_141150:

Is, is the compensation in those labs,

400

is it, is it done on a per case basis?

401

Or, I mean, is there…

402

I'm just trying to think about kind

of like, in, in my, in my mind, what

403

you're describing feels a lot like Uber.

404

It feels

405

eric_1_05-05-2026_141149: Yeah.

406

dr--andy-roark-_1_05-05-2026_141150:

you just, you flip on the meter and

407

you drive for, you know, as long as

you kinda want, and the number of trips

408

you make or the number of miles you

drive is kinda what you get paid for,

409

and you, you flip the meter back off.

410

Is that kind of where pathology is

going into this kind of Uber-ification

411

of, of this part of medicine?

412

eric_1_05-05-2026_141149: Yeah, so the

old model was most people were on salary,

413

and you could kind of think of it almost

like as a form of pro sale, where if you

414

dr--andy-roark-_1_05-05-2026_141150: Yeah.

415

eric_1_05-05-2026_141149: expected

benchmark, you would often

416

get a pay-by-case bonus above.

417

they're, they're pretty modest.

418

if you had to guess, what

would you say the pay-by-case

419

bonus is at most of these labs?

420

dr--andy-roark-_1_05-05-2026_141150:

I, I have, I have no idea.

421

I'm trying to think, you know?

422

So he's doing the cytology slides,

it's, you know, $300 to the clients.

423

I'm like, I don't know, 20 or $30?

424

Uh.

425

eric_1_05-05-2026_141149: So there's a few

places where you get that, but a lot of

426

them it's honestly about eight, 10, $12.

427

dr--andy-roark-_1_05-05-2026_141150: Okay.

428

eric_1_05-05-2026_141149:

Europe, it's as low as $5 a case.

429

dr--andy-roark-_1_05-05-2026_141150: Okay.

430

eric_1_05-05-2026_141149: you know, you

can see how people feel pretty squeezed.

431

the s- the salaries aren't great.

432

The pay-by-case isn't great.

433

And so some of the people who do wanna

make a lot of money, and we all have

434

huge student debt now, they end up just

reading these astronomical case volumes.

435

dr--andy-roark-_1_05-05-2026_141150: Wow.

436

eric_1_05-05-2026_141149: the number

of slides you're supposed to read for

437

cytology in human medicine in the US

is legally capped at 100 slides a day.

438

And I would say people in the big labs are

often reading multiples of that every day.

439

dr--andy-roark-_1_05-05-2026_141150: Wow.

440

eric_1_05-05-2026_141149:

it, it's very high volume.

441

And yeah, the Uberification is

definitely a thing that started to

442

frankly come with, digital cytology

and more of kind of freelance model.

443

And, and I own up to, you know,

I started Lacuna Diagnostics, and

444

we relied a lot on contractors, so

445

dr--andy-roark-_1_05-05-2026_141150: Yeah.

446

eric_1_05-05-2026_141149:

our roads have diverged, even

447

if the models are similar on

448

dr--andy-roark-_1_05-05-2026_141150: Yeah.

449

If it-- I feel like if you hadn't done

it, someone else would've done it.

450

Everybody's looking from a, you know,

an entrepreneur standpoint, everybody's

451

looking at how to, at how to scale,

you know, the program that they do.

452

Everybody is looking for how to have

this model where you can, you know,

453

have a flexible workforce that can work

as much as they want, and then you can

454

keep them as independent contractors.

455

I don't think that there was

anything shockingly novel.

456

No offense, but I don't feel like

there's anything shockingly novel about

457

this that like, you know what I mean?

458

Other, other people wouldn't

459

eric_1_05-05-2026_141149:

Technology was the novel piece.

460

Yeah, the idea-- I mean, like the brick

and mortar era with glass slides, there

461

were contractors at all the big labs.

462

So

463

dr--andy-roark-_1_05-05-2026_141150: Yeah.

464

eric_1_05-05-2026_141149: with that.

465

It made it maybe change the

volume and the pace and the

466

kinetics and some of the policies.

467

But yeah, and part of this is, you

know, I did-- after vet school,

468

I did a rotating internship.

469

I worked ER relief for a number of years.

470

I saw how impactful it could

be when you couldn't get those

471

cytology results for a few days.

472

So

473

dr--andy-roark-_1_05-05-2026_141150: Yeah.

474

eric_1_05-05-2026_141149: my driving force

for going towards digital was to really

475

get those results when they matter in a

matter of hours rather than days or weeks.

476

dr--andy-roark-_1_05-05-2026_141150:

there's something like if you said, "Oh,

477

Henry Ford and his assembly line, look

at all the, the pain that he brought."

478

And you go, "Well, it's a tool."

479

Like this is a, it's a tool.

480

It's a way of doing things.

481

It's wildly efficient.

482

It, it created a lot of opportunities.

483

And there are some people who apply

that tool in a, less than ideal

484

way or there's, you know, there's

definitely people who have worked on

485

inch- assembly lines and, and been

miserable and had their health and

486

safety discounted and things like that.

487

And so to me, it's kind of the same thing.

488

It, it's, it's there's ways

of doing things and, you know,

489

how do people apply those?

490

I, I think the idea, you know, you

look at the radiologists and they're,

491

they're doing, they're doing quite well.

492

I, I think that that's an important

point in sort of how much, how much

493

responsibility you take, how do

494

you, you know, how do you, how

do you feel about how things sort

495

of, things sort of shook out,

496

eric_1_05-05-2026_141149: And to extend

that further, we now pine for the

497

days when everyone worked these good

manufacturing jobs on the factory floor.

498

dr--andy-roark-_1_05-05-2026_141150: Sure.

499

eric_1_05-05-2026_141149: towards

the knowledge work economy, we have a

500

different set of problems, but we're

longing for this halcyon era that may

501

have either never existed or not been

as good as we remember it with the

502

rose-colored glasses looking backwards.

503

dr--andy-roark-_1_05-05-2026_141150:

Yeah, I, I, I think I, I may be a, a bit

504

naive in some ways, but I, I really…

505

I have this idea that I, I would like

to see, I would like to see industry

506

in America go towards this, not just

maximizing revenue, but, but having an

507

optimization of revenue, like purpose,

meaning like the, the good you do

508

in the world, and then environment,

which is how you treat your people.

509

And so I just think wouldn't it be

great if you could have a profitable

510

business that had a purpose and

that you treated your people well?

511

Like, is that too much to ask?

512

And again, if, if you measure all of your

success based on your quarterly earnings

513

and your stock performance, it's, it's,

it, it contradicts exactly what I'm asking

514

for, but I, I miss, I miss those days

and I, you know, I, I would like to see,

515

I would like to see more companies kind

of focusing on balancing those things.

516

eric_1_05-05-2026_141149: Absolutely.

517

I, I may also be naive, but I,

I was taught and I still believe

518

good medicine is good business.

519

And I think that there should be room

for us to turn a profit with a good

520

patient-focused, client-focused service.

521

And I, I do think we've

tilted more towards one side.

522

but part of why that happened is for

a long time veterinarians weren't

523

particularly good business people.

524

We didn't run things well, and it was

rife for having people with MBAs to come

525

in and tell us how to lose less money.

526

dr--andy-roark-_1_05-05-2026_141150: Yeah.

527

No, I, I, I, I do, I do think

that we have made our bed to some

528

degree for that, for that reason.

529

Again, I, I love, I love being

a veterinarian and I love vets.

530

I, I think our-- we had a pretty strong

anti-business mindset, and I think it

531

left the door open for, for others to come

in and, and, and grab up opportunities

532

that we were, that we were leaving,

hoping no one else would pick up.

533

And so I think that that's, I think that

that's something, that's something I wish

534

we could go back and change, but I think

it sort of-- it is, it is where it is.

535

Eric, it's five years in the future,

and the pendulum has swung back the

536

other way, and pathology is great.

537

What does that look like?

538

What has happened?

539

What, what-- Just leave, leave

me on a happy note of what

540

happens if this is great?

541

What happens if this

really goes well from here?

542

eric_1_05-05-2026_141149: Sure.

543

Well, I'll take that into a couple pieces.

544

So one, on the volume and economic

piece, counter to a lot of my concerns

545

is something called Jevons paradox,

546

dr--andy-roark-_1_05-05-2026_141150: Okay.

547

Okay.

548

eric_1_05-05-2026_141149: an economist

in the industrial era in Britain.

549

And he noticed that when, James Watt

came out with his more efficient steam

550

engine, a lot of these things that

were consumed in terms of coal and

551

everything else from going down because

it was more efficient, they shot way up.

552

And it's since been applied to mean

any time a resource or technology makes

553

something cheaper or more efficient, it

tends to increase utilization of that.

554

So we've seen that even in lab medicine.

555

So in the old days, people did

a CBC by actually doing all

556

the different assays by hand.

557

They would be doing a hemocytometer count,

things that would be unimaginable today.

558

But you really can't do 40, 50, 60 CBCs

a day even in a small lab doing that.

559

So as we started to get these high-powered

instruments, both like a little point

560

of care thing like a Procyte or bigger

systems like a Sysmex analyzer or

561

an ADVIA hematology analyzer, the

volumes exploded and went way up.

562

So the business side may

work itself in terms of that.

563

In terms of how AI is gonna be

applied to our workflows, the

564

positive version of this is what we're

starting to see in human medicine.

565

So they have a lot of strict

regulations that we do not at all

566

have in human medicine, and so they

have to treat AI and other forms

567

of software as a medical device.

568

So there's a lot of scrutiny in

there, and as a result, their model

569

has not been automated diagnosis

straight to the end user clinician.

570

it is more of an assistive capacity.

571

So they do things like there are different

softwares that predict a cancer grade.

572

For prostate cancer, it'll give

you a Gleason score, and it'll

573

show you its work, and the

pathologist can accept or reject it.

574

So it makes things a little faster, but

it's also a person's in the loop calling

575

dr--andy-roark-_1_05-05-2026_141150: Yeah.

576

eric_1_05-05-2026_141149: had a bunch

of tools like that and it was very

577

modular, I would actually love that.

578

That would be great to be able to

be like, "Hey, it caught something

579

I missed on s- on my fourth case

of the day that was really tricky."

580

Some places are using it to triage

triage up the harder cases higher in

581

the queue earlier in the day, so that

by the time you're burned out at 4:00,

582

5:00, 6:00 at night, you're reading

your lipomas and your dumb stuff.

583

But the hard, challenging

cases, the suspected cancer

584

cases come early in the day.

585

I think that's a great use.

586

of the uses we're seeing at some of

the companies for radiology are great.

587

They're applying vertebral heart scores

or their automated measuring of TPLO

588

things, but they're not saying, "This

dog has a stage two osteosarcoma.

589

You should euthanize it."

590

They're very applied tools that are gonna

be saving time but also have the ability

591

to be changed or accepted or rejected

by the clinician or the radiologist.

592

So I think those two things.

593

So economics may sort itself out, and I

think that if we start using this in a

594

fashion where it's very much augmenting

human intelligence, I think that that

595

could make our role more interesting.

596

We could be focusing more on the harder

cases and not as much rote nonsense.

597

I do worry that the piece in veterinary

medicine is gonna focus very much on the

598

point of care, so people aren't even going

to be sending those cases to the lab.

599

That's where the analogy could

break down from previous things.

600

You know, there-- everyone

says this time's different.

601

Well, this time may be different with AI.

602

I don't know that it will

be, but I worry it might

603

dr--andy-roark-_1_05-05-2026_141150: Yeah.

604

Well, Dr.

605

Eric Fish, thanks so much for being

here and talking through this with me.

606

Where can people find you online?

607

eric_1_05-05-2026_141149: obviously

on LinkedIn I post things there.

608

A lot of it is cross-posts for

my newsletter, Science Great

609

and Small, which you can find.

610

We've got in the show notes here.

611

So I, I write a lot about AI,

but that's not the only thing.

612

I break down a lot of different studies.

613

I write about some of the big macro

trends in veterinary medicine, for both

614

the clinical side and the pathology side.

615

So people enjoy it.

616

dr--andy-roark-_1_05-05-2026_141150:

Excellent.

617

Thank you so much for all you do.

618

Guys, thanks for tuning in, everybody.

619

Take care of yourselves, gang.

620

eric_1_05-05-2026_141149: Thanks a lot

621

Speaker 2: And that's our episode.

622

Thanks to, Dr.

623

Eric Fish for being here.

624

Guys, thanks for tuning in.

625

I hope this was interesting for you.

626

as always, if you, if you like the

episode, go ahead and do the things

627

that people do when they like podcasts.

628

like share, subscribe.

629

Send it to a friend.

630

All of those of Right.

631

A fan mail to me.

632

I, I, I would take that, that

would be a, that would be a

633

thing that I would be okay with.

634

Anyway, guys, I'm messing around.

635

Have a wonderful day.

636

Thanks for tuning in and listening.

637

I'll talk to you later on.