Episode 416

full
Published on:

18th Nov 2026

Productivity

Why has productivity growth slowed across advanced economies, even as technology makes us more capable? And what happens when the value we create becomes harder to measure?

In this episode of the Cognitive Engineering Podcast, Fraser, Peter and Nick explore the UK's productivity puzzle, the limitations of economic statistics, and the impact of technology and AI. The discussion also questions whether some jobs create genuine value or simply add bureaucracy, processes and management overhead. Does the productivity puzzle reflects a real economic problem, flawed measurement, or something else altogether?

Links:

UK productivity 'a bit stronger than thought' (Reuters)

The Productivity Institute covers the UK's productivity problem.

David Graeber's Bullshit Jobs.

The World Management Survey.

The Solow Paradox

Leprechaun economics



This podcast uses the following third-party services for analysis:

Podtrac - https://analytics.podtrac.com/privacy-policy-gdrp
Transcript
Speaker A:

Hello and welcome to the Cognitive Engineering Podcast brought to you by Aleph Insights and produced by me, Fraser McGruer.

I'm here with Peter Coghill and Nick Hare of Aleph Insights.

On this podcast we look at a wide range of topics from an analytical viewpoint.

And today we're discussing productivity.

Nick.

Productivity.

Yes.

Speaker B:

Well, some good news for the country of Great Britain.

Speaker A:

Excellent.

Speaker B:

It turns out we have had a sort of fairly, fairly sizable revision to the UK's productivity statistics over the last 10 years.

So listeners should strap in and get ready for some exciting stats chat.

Speaker A:

Good.

Speaker B:

You know how we used to thinking of Britain as the sick man of Europe and we're just a bunch of sort of lazy malingerers and, you know, all we do is work in call centres and go shopping in Greggs.

None of that's true.

None of that's true.

Apparently we're like, we're a nation of busy bees constantly churning out output.

Obviously neither of these are totally true.

Let's dive into it.

What do we mean by productivity?

Well, economists have got a perfectly precise definition of that.

Speaker A:

Okay, well, hold on.

Speaker B:

So the news is, I mean, productivity is higher than we thought, right?

We're not as much of the sick man of Europe, the sort of British disease of low productivity, which we've been kind of people have been talking about for decades.

Speaker A:

Decades.

Speaker B:

So productivity has got this precise definition.

It is gross value added per hour worked, right?

Which means you take the value of all the.

All the value we added as a country.

And by value added, I mean if you spend an hour working on 90 pounds of metal to turn it into 100 pounds of tin trays, you've added £10 of value.

So we don't look at the value of the tin tray you produced, which is £100, we just look at the extra added value.

Now that happens to be the same as gdp, so.

So we tend to use GDP as the sort of amount we produced.

That's how GDP is calculated, basically.

And then we take the number of hours we worked and we divide one by the other.

So it's the amount of output we produced per hour that we worked.

And then of course, all of that gets adjusted by inflation.

And so we can track productivity over time and so on.

So what happened?

Did we suddenly discover that we were producing more stuff than we thought?

No, it was the bad kind of productivity increase where instead of the numerator getting higher, the denominator just got low.

It turns out we're just skivers and we were working far fewer hours than we thought.

And so it turns out we lied on our timesheets.

Well, more or less, because in fact, it's not exactly lying.

But we used to use the ons.

Always used to use this thing called the labor force survey.

And it turns out that there was.

I mean, there's all kinds of problems with it, but all the, all the problems with it actually a bias hit towards making it look like we work more hours.

So, for example, some people didn't fill it in because they were on holiday.

So you're already losing the people who are on holiday.

Speaker C:

Right.

Speaker B:

Which, which is going to tell you something important.

But, but also people would, would often just routinely just report their contracted hours.

So if you're contracted for 38 hours a week, you might just write, oh, I worked for 38 hours this week.

Even though you might have spent, you know, the whole of Monday in bed with hangover, you know, and Wednesday was.

You took a half day off to, I don't know, to go and play in a five or six.

Speaker C:

You're working from home and you're skiving off, walking the dog most of the day.

Speaker B:

That's right.

We need, we, we really would like to know how much stuff on average people are able to produce in an hour at work.

And if we don't know how long they're at work for, we cannot tell how productive they are.

Speaker A:

Right.

Speaker B:

So anyway, it turns out they've now got this compositional measure where they, they use a number of different sources now, number of different ways of estimating it.

And it turns out we've just been working for less time than we thought.

So we're not, we're actually.

It's bumped us up a little bit now.

This is kind of output is.

% higher than it was in:

We now think it's 40% higher than.

Speaker A:

And what do we think it was before?

Speaker B:

Sorry, 34% higher.

So it's not.

Yeah, it's, it's a, it's a matter of a few percent.

Yeah, it's not bad compounded.

You know, if that kept going, if we kept growing faster than.

Speaker C:

If we could just pull our socks up, get back to work.

Speaker B:

Exactly.

Speaker C:

Imagine what we could achieve.

Speaker B:

Well, exactly.

So, however, this has not solved the big problem.

Right.

Which is, yes, there's still a productivity problem.

CD, productivity growth since:

And no one knows precisely why.

Right.

there's a financial crisis in:

It is like it's very, very visible in a long term chart.

And the UK table.

We beat ourselves up about productivity but as far as we can tell, we're, we're surprisingly higher, quite a bit higher than Japan.

Speaker A:

Okay.

Speaker B:

Possibly tells you something about Japan's work culture.

You know, that they spend too much time at work, not really doing anything.

But we're above Italy, we're above Canada, we are quite a, quite a long way below Germany and the US and sad, sad to say, quite a bit below France.

Speaker A:

What?

Speaker B:

Yeah, yeah, the, the famously facts and lazy French who spend most of their afternoon drunk.

Speaker C:

When they are.

Speaker B:

I mean, I think there may well be something to that as well.

Yeah, but.

And we're actually, we are pretty much at the average for the euro, euro area.

Speaker C:

So I take that as an argument for longer lunches and afternoons off.

Speaker B:

I think, I think we might well come on to that.

Speaker A:

I, I take this as a kind of an affirmation of the British way of life actually, which is this sort of, you know, sensible middle approach to things.

Let's not be too ideal.

Speaker B:

Don't want to try too hard.

Yeah, try too hard.

Why do we?

The word try hard is an insult.

I bet it's a compliment in Germany.

Speaker A:

Yeah, exactly, exactly.

Yeah, yeah, yeah.

As in so many things, you know, we've, we've a happy medium, basically.

A sensible compromise.

Speaker B:

Anyway, I think it's great, I love that, you know Those sort of 18th century travel writers where whichever country they're from, they always end up concluding that they've got it exactly right and foreigners are either two of one thing or too much of the other thing.

But anyone who's ever visited Britain must surely conclude we've got everything right.

Exactly right.

So we've got what makes productivity go up.

It's quite straightforward actually.

Better technology, which as the way that economists use the term technology basically means anything that makes you able to make more stuff with the same inputs or more basically more productive capital.

So you just, you know, you've got one person with one machine, you know, might be able to produce twice the amount if they had two machines.

Speaker A:

Sure.

And so sorry, you might already.

There might be an existing answer out there and I'm not asking anything new and you might know this already, Nick, but so is the question if we're talking about linear improvements.

Right.

Which is that data that you mentioned suggests that that's the case.

Right?

Speaker B:

Yeah.

Speaker A:

Because then suddenly in:

Right.

uggests something happened in:

Is it causal correlated?

I don't know.

But anyway, I was just wondering, is this a question of.

And I think we've talked about this on a podcast before about, you know, diminishing returns on like, you know, the hundred meters sort of athlete that's, that's got, there's a point where you just can't have that linear improvement and it's just almost impossible.

Speaker C:

Yeah.

Speaker B:

So have we, have we actually eaten all the economic low hanging fruit?

Yes, there is, there is that.

I mean, I think that's one of the things people discuss, I presume.

Yeah, yeah.

That, that actually kind of a lot of the gains from the last, let's say 50 years, 100 years were things that were kind of were to do with like distributional effects, to do with globalization, the internationalization of capital, the fact that you can go, oh well now China's appeared, so we can, now we, you know, we can offshore all of the cheap, all of the kind of low productivity manufacturing stuff and the US can then just keep doing more and more of the kind of, you know, the high productivity things like designing logos or whatever it is they do, someone else can make the phones, which is relatively low productivity.

That is definitely a theory that people have advanced.

I should say there's no consensus at all.

It's still considered to be a bit of a mystery, but, but yeah, the idea, the idea that, you know, you've kind of used up all of the cheap productivity improvements and because of internationalization we got to a point where that's no longer a constraint.

We've now removed that constraint altogether.

And so, you know, productivity growth is just going to be.

We're now up against the frontier, you know, and we're having to generate, generate that through investment and through, and through technological improvements.

Speaker C:

I mean, I think a way to think about it is like a good big juicy piece of low hanging fruit was deciding on a standard for shipping containers.

Speaker A:

Yeah.

Speaker C:

Which meant that dock workers could be paired right back and any given dock worker could become 10 times more productive because they can move 10 times as much stuff.

Sure.

But like what's kind of shipping containers version 2.

There isn't one solved.

There's nothing now.

Speaker B:

Yeah, we've used up China, we've used up shipping containers.

I mean that's the, this is the kind of nonsense.

There's this sort of thing those.

It's productivity and I suppose why it's quite hard to sort of deal with is productivity is just loads and loads, thousands and thousands of little tiny decisions like that.

Speaker A:

Okay.

Speaker B:

It's like thousands of little decisions, each one of which makes someone somewhere a tiny bit more productive.

Speaker C:

But I also think there's this cultural things like safety culture eats in the other direction because if you know that dock worker could be more productive than they are if they didn't have to follow those safety rules and things.

Speaker A:

Although I wonder if long term it does help productivity.

I could imagine the short because they.

Speaker B:

Don't die so much.

Speaker A:

Yeah, you have to recruit new ones because last lot got slaughtered or.

Yeah, yeah.

Speaker B:

Anyway people got crushed by a tea crate.

Speaker A:

Yeah, exactly.

So could we sum this up where we are at the moment either Nick or Psych.

Speaker B:

Basically the good news about UK productivity hasn't really changed the big picture, which is the UK is not super productive but also it hasn't solved the big mystery of where our productivity growth went to.

So yeah, I mean just a couple of other things before we get on to.

Before we get on to I think probably the interesting bit which is actually like well, are we measuring the wrong thing?

But just so the other things that people tend to blame on the UK having relatively low productivity is just quite low capital investment.

So you know, we tend not to invest as much as other countries in the oecd.

What Peter was talking about, there's sort of infrastructure and planning constraints and things so transport congestion, commercial property is quite expensive.

It's hard to expand in productive locations.

You know, just literally all the kind of legal and planning issues that you have to do.

And then we have relatively bad management in the UK so far as we can tell from there's a thing called the World Management Survey which these guys, Bloom and van Reinen where they measure 18 things which are about kind of operations and monitoring, target setting, people and incentives.

Right.

This just measures that.

It doesn't tell you whether those things are actually particularly important but.

But it does correlate with economy wide productivity and American firms score better than European firms on all of that stuff.

Unlike these visible is this thing being well managed.

But the UK has a long tail apparently of badly managed firms.

So we have the UK has a lot of bad management, more than it should.

And some part of that is to do with relatively weak competition, allegedly.

But also like inherited management.

So poor talent management, I guess so.

So, yeah, I thought it was quite interesting.

All right, Quite an interesting observation.

I don't think it tells us a huge amount, but it chimes with our sense of.

When I think of a British firm, I think of are you being served?

Yeah, it's that kind of think of Reggie Perrin, you know, our kind of standard idea of, of what a British company is like is inefficient and.

And an American company, of course, ruthlessly efficient.

Ruthlessly efficient, yes.

Speaker A:

But isn't there something wonderful about that?

Yeah, I hate the idea of this sort of awfully efficient company.

I know which one I'd rather be in.

Speaker B:

I think, I think there is.

Yeah, there is something to that.

So.

And one final thing before we get on to diving into a bit more of the kind of the, the measurement side of this jobs.

obably heard of David Graber,:

This, this has been.

It's sort of challenged in the serious literature, to be honest, but I think there's got to be something to it.

The fact that this spoke to so many people, but I think he identified this increasing category of kind of classes of jobs that don't actually do anything and inexplicably exist.

So he's got this kind of five way taxonomy where you have flunkies whose only job is to make someone else look important.

So having a personal assistant or something, goons who do adversarial work that wouldn't be needed if competitors didn't do it.

So things like search engine optimization, telemarketing, those kinds of things.

Duct tapers, they fix problems that shouldn't be there in the first place.

Box tickers who produce paperwork or procedures that create the appearance of useful activity and taskmasters.

These are people who supervise people who don't need supervising or who invent unnecessary work.

Dunno, that sounds familiar to me.

Speaker C:

We've met a few.

Speaker B:

Yeah, we've employed a few.

Speaker C:

Yeah, yeah.

Speaker B:

I mean, I think as I said, it's sort of challenged the extent to which this is an issue.

But it is definitely the case that if you as a company have got a team of 10 people who are a bit chaotic but highly productive, you then employ five managers to give the appearance of better management.

But if they're not actually doing anything, then all you've done is reduced your productivity by a third, you've pushed up your wage bill and, and if you're adding jobs that are you.

You're only adding because you sort of feel like you kind of need them rather than because you.

They're actually doing it.

Yeah.

So.

So I. I think.

I think.

But it's interesting that.

That does feel to me sort of connected to the.

The question of.

Of unpro.

Of sort of unproductive firms.

Speaker C:

Yep.

Speaker A:

Okay.

Speaker C:

Yeah.

Speaker B:

But I mean, so what's.

I think the truck.

Speaker A:

One.

Speaker B:

One of the reasons why this is so hard is because it's actually really hard to measure the thing we care about.

Speaker A:

Peter, you got something on this?

Speaker C:

Yeah.

So I think.

Yeah.

Measuring the right thing and measuring and choosing what to measure.

I think key questions.

There's an interesting thing.

It describes like, it's subtle.

What seemed like subtle changes of measure have drastic effects.

Have you come across the leprechaun economics?

Sadly not, no.

So in:

What report?

That's what it reported.

Speaker B:

They all suddenly woke up.

Speaker C:

They all suddenly.

They all suddenly woke up.

One quarter quadruple espresso and got to work.

And that year, January 1st.

And then suddenly, you know, the 26%.

No, they weren't.

What happened was they changed how they measured their gross domestic product.

And they went from a. I'm not fully understand it.

Maybe Nick can explain, but they went from a traditional measure to what they called the gross national income.

Speaker B:

Yeah.

Speaker C:

Measure which basically took into account things sort of like took into account profits and profits made overseas as well as just their sort of internal product would.

Speaker A:

Kind of make sense.

Speaker B:

And because they have these kind of international firms there because of tax breaks.

Speaker C:

Yeah.

Speaker B:

You can see.

Yeah.

Speaker C:

So that's island.

And that's a sort of change in how they measured things.

Just highlight how a change in measurement can have a big effect on what you're actually.

What you actually see.

Speaker B:

Well, the Italians did a similar stunt.

Have you heard of Il Sorpasso?

Speaker A:

No.

Speaker B:

It means the surpassing.

great national moment in the:

I think it was the:

Speaker A:

All the layers.

Did they overtake someone?

Speaker B:

They overtook us.

Speaker A:

Yeah.

Speaker B:

Their GDP overtook us.

And you know, do you know how.

Because.

Because they decided that organized crime should be included in gdp.

Oh, wow.

Speaker A:

Yeah.

Speaker B:

I mean, hey, presto, free wealth.

Speaker A:

I mean, you can kind of see the rationale behind that.

Speaker B:

But yeah, it's kind of completely justifiable decision, but it is nevertheless very funny.

And it's the fact that they.

The fact that they were so cock a Hoop about it as well.

Speaker A:

Yeah, yeah, yeah, yeah, yeah.

Okay.

Anything more on measurement issues or have we got somewhere else to go?

Speaker C:

Yes, I think we should talk about measure enlightenment with respect to how things are changing, particularly with AI.

Okay, so just to highlight this point, the economic benefit of technology is actually quite difficult to spot in headline figures.

Take for example, the Solo paradox.

This was described in:

Robert Solow, where he was trying to find within the figures where all the benefit from computer technology.

So computerization was a big thing.

You know, businesses were building big data centers and desktops were appearing and people were moving from spreadsheets to spreadsheets and everything.

So computerization was really kicking off in the sort of late 80s to 90s.

But a good quote is you can see the computer age everywhere except in the productivity statistics.

So you couldn't see like all this investment going into these computers and things, and people were obviously kind of better at their jobs, able to do more, but it didn't appear in the, in the productivity stats.

Speaker A:

And so by that do you mean it didn't appear in the productivity stats because productivity didn't go up or jump as you might expect, the rate of growth, or that the rate of growth went up, but you just somehow couldn't quite see how that was going through.

You can itemize it or identify it,.

Speaker C:

It's a little bit of both.

Another historical example is we're doing electrification industry.

So before electrification, factories were driven by steam boilers, before that, water and other things.

But steam movement from steam to electric had lots and lots of benefits in terms of the technologies you could apply.

You could have smaller, more distributed motors and things on machines rather than relaying mechanical power throughout the factory, which is harder to maintain, etc.

So basically, huge, huge potential benefit, but again, a long delay between all of that happening.

Productivity looking like it's getting better.

You know, observationally, you can see people churning out more things, but in the stats doesn't appear, it's just not reflecting.

Speaker A:

You can imagine why there might be a lag.

Speaker C:

But anyway, sorry, yeah, essentially it's lots of different effects, but for going back to the computing, it's because you require any electrification.

There's a sort of network effect that kicks off.

Once you've got enough things in the network which are doing things the same way, then you start seeing the real benefit from it.

So, for example, if I electrify my factory and you and I'm making parts that you are then putting into automobiles, but you haven't Electrified, yet I can spit out hundreds, ten times more parts per day.

You can't then accommodate that, so the parts sell cheaper or I produce less and I start producing something else.

So it's.

Unless until we're all electrified or all computerized, you don't get the overall benefit from it.

Speaker B:

I think there is another, more fundamental thing about how we measure the unit of output for a lot of the things that are not just a kettle or a spoon.

Going back to accountants with spreadsheets, right?

I mean, when we look at how much sort of accountancy services people get produced, we simply cannot measure things like quantity, numbers of accounts audited, or the quality of accountancy that's been provided.

Like, we measure that just by looking at what people typically things like, well, what, what's an hour of an accountant cost?

How much, how much do people pay for an hour of accountancy services in inverted commas, how much do people pay for, you know, some computer programmer output?

And we adjust that in real terms, we go, well, we're paying, paying this guy for an hour of his output.

And now computer programmers more valuable because we're paying them in real terms twice the amount that we did 10 years ago.

But we don't know.

No one obviously can or does attempt to measure the quality on a national scale, attempt to measure the actual quantity or quality of their actual output.

How many computer programs have you written?

How many lines of code have you written?

How reliable are they?

No one measures that.

So these productivity improvements that affect that kind of thing, those invisible quality improvements are just not going to appear.

We don't have any means of capturing them.

So actually, things and quality as well, typically the way that we work out inflation is by taking a single thing and tracking it over time, saying, well, how much is a computer now compared to a computer 20 years ago?

And computers have got much cheaper.

So we go, brilliant, right, we're richer in a sense, because computers, we can now buy more computers with every pound that we earn.

But in fact, it's been a plummet because if you look at the cost per gigaflop or something, it's been hundreds of times, thousands of times, tens of thousand times improvement in the quality of that thing.

And it's not appearing anywhere.

Speaker A:

So I see what you're saying.

It's actually not that Britain has a productivity problem, it's just that.

It's just.

It's not reflecting the massive quality we're feeling everywhere.

Speaker B:

phone, like a mobile phone in:

That was as powerful as a mobile phone.

But.

But if there was, it would have cost tens of millions of pounds.

Speaker C:

Right.

Speaker B:

nse, like we should say we're:

And then quite apart from the fact that all of what you can do on it are things like look, look up any fact ever, which you people used to have to spend thousands of pounds on encyclopedias to do that.

Find your way anywhere without a map, where instead of buying a map, I mean, there's just the things you can do now which don't appear anywhere because they're free, they're not gonna appear anywhere.

Speaker C:

And now you've got Claude or ChatGPT sitting on, acting as your personal assistant, who you would have to employ it sort of 10, 20, 30,000 pounds a year, but now you spend 20 pounds a month on it.

Speaker B:

Yeah, yeah, but if we had an employee that was as.

As good at Claude, we'd probably have to spend a lot more than that.

Yeah, but I mean, that's a really good example where I think, yeah, I used.

I use Chat GPT a lot, I think in an appropriate way, but.

But of course everyone's going to say that, but the point is that I spend £20amonth and I would don't tell Sam Altman this, but I'd be willing to spend a lot more than that.

Speaker A:

Yeah, yeah.

Speaker B:

Just because of how helpful.

Speaker A:

Yeah.

nking about, I remember circa:

And it will tell you different ways to get to Hampshire and they post it to you, how you do that,.

Speaker B:

That's such a glorious in between, isn't it, of two worlds.

They probably had a computer that did that.

Yeah.

And so they would what, read it out over the.

Speaker A:

No, no, no.

So they, you know, I guess you'd give two postcodes, let's say, and then I think it was the AA would come back to you like a few days later and you'd receive, you know, in the post, some pages presumably have.

Speaker B:

A little route map.

Speaker A:

Yeah, yeah, yeah.

As well.

Yeah, yeah.

And like waypoints and all that sort of thing and alternatives as well.

Speaker B:

And I mean, any recollection of how much this was?

Probably we're talking about a fiver or.

Speaker A:

Yeah, I think it's all like five, ten quid, something like that.

Yeah, yeah, yeah, yeah.

You know, and of course, we all know, you know how one might do that now.

Yeah.

Speaker B:

So, I mean, on the average, let's say you might use a route map once a day, twice a day to get 20 quid a day, you know, £100 a week, £500amonth of just free stuff which doesn't anywhere appear in our product.

So if productivity improvements have been concentrated, which seems plausible to me, in the production of these things which either don't appear in GDP or if they do appear, are massively undervalued, then, you know, maybe that's where our productivity growth has gone.

It's just.

It's there, but it's disappeared.

It's not being captured anywhere.

picious because, hey, presto,:

So it's like, well, you know, has productivity just productivity growth been disappearing into an invisible.

Invisible, you know, but nonetheless real.

Speaker A:

Yeah.

Speaker B:

Set of activity.

Speaker A:

But anyway, I don't want to go too much down this, but then the question is, you know, Britain versus other places.

Because you would expect a similar kind of growth across everywhere, or lack of, you know, what you've just said, to be reflected in everyone's day.

Speaker B:

Well, no, but that is so.

So everywhere you look, you see the same king.

So it is everywhere.

Speaker A:

Yes.

Okay, cool.

Speaker B:

I mean, I, I have to say, people, actual economists have looked at that and allegedly it is a contributor.

I wonder, I personally wonder if we're not actually hugely underestimating that.

But yeah, sorry, Peter.

Speaker A:

Peter.

I think you started talking.

I think we want to move into AI.

If I'm right, and you were going to talk about that a little bit,.

Speaker C:

We bang on about AI so much.

I just sort of had a realization that.

Do we.

Okay.

Speaker A:

I don't know if it's relevant.

We do, I guess.

Speaker C:

I mean, I think it is relevant because AI is poised to completely radicalize how we produce.

Taking a slightly AI Boomer perspective, there's a chance that AI will simply replace all human knowledge work.

So aside from the job displacement problem, now that you've got now, you can just sort of print.

They just fab new AIs to do more human work.

You can just.

You can just multiply the amount of work that gets done.

How do we then measure this?

Because we you'll have these effects where the quality, potentially you can just decide what quality you want the work to get and how much of it.

So the measures of GDP mean much less.

Also national, the sort of national gdp.

If the AI tech giants are cornering all of the productivity, does stuff that is done in the UK still count as Ukraine productivity or is it?

Speaker B:

Yeah, well I think that's a perennial problem, isn't it?

Yeah, if you sort of value valuing a brand which exists in the UK as a brand and has its headquarters in the uk, but actually where all the activity takes place overseas or something, it's like, well, where is the value actually being added?

Yeah, I think, I mean, I think this is going to compound the problem we've had of not being able to capture that kind of productivity increase.

Because when I think of, well, you know, what does it mean for me?

It means that I can do in a day when I think, well, I want to produce a piece of code which does something and it would have taken me a month, now I can get it done in a couple of hours.

That's a massive productivity increase which is not really going to appear anywhere because the kind of thing that I'm producing at the end of the day is still going to look like it's a technical report or some project output that someone is paying for.

You know, it's not an.

In theory I could do, I could maybe do 10 of those instead of previously I would have just been able to do one.

But it, but it.

In tactically that isn't.

All of the other stuff means that that productivity improvement, like the, the quality isn't, doesn't appear anywhere or.

Speaker C:

Yeah, yeah, but there's, there's arguments against counting artifacts as the unit of productivity because, you know, like human created things.

Let's just assume that the quality will kind of stay about the same.

But the AI is kind of.

All the emails that AIs are going to send to each other and intermediate documents are going to produce for each other.

It would indicate a huge boom in productivity.

But is that just the slop or, you know, is that just like all that intermediate stuff?

Is that just like the bits in the computer?

We don't count those.

We wouldn't dream of counting them.

Speaker A:

Yeah.

Speaker C:

What, where do we put, where do we do the counting?

Is it on the outputs or is it in the intermediate stuff?

And what is an intermediate versus an output?

Speaker B:

Yeah, I mean it is no different in principle, you know, having, having AI.

If we had AI enhancements to a spoon factory you know, where now we used to need to have 100 people making spoons and now we have one guy making the same amount of spoons because all he's got to do is turn the machines on.

That is a hundredfold increase in productivity.

That's literally, that's a canonical example of how you get a more productive man.

His hour is now producing 100 times more spoons.

Well, you know, AI for some industries for knowledge work is going to have a similar sort of effect.

But we cannot count knowledge, we can't count the output.

We don't know how to measure the output of knowledge work in a way that makes it comparable to the situation with spoons where we can just count the number of spoons there.

Is that so?

I mean, I think so.

I think, yeah, that's going to be one of the issues and it certainly has been remarked on and discussed that, well, where's the huge productivity gains that we're supposed to be getting from AI?

You know, these things have been around for two or three years now.

If the, you know, the hysterical claims that people make about being 10 times or 100 times more efficient of being true, where is it?

Well, it may be is there, but it's just invisible.

Speaker C:

Yeah.

Speaker A:

So interesting.

So this is not, I mean, in one sense you could call that a measurement or metrics problem, but actually it's much more fundamental than that.

It's the sort of, it's much higher level than that.

Right.

Speaker B:

Which is we don't know what we're counting, so we don't know how much more of it there is.

Yeah, exactly.

How much more have I made per hour?

The question is almost meaningless.

Speaker A:

Yes, the wrong question.

Speaker B:

And I mean.

And so sorry.

It probably brings us worth discussing the.

Going right back to the revision to the productivity statistics that.

The other thing which I think is becoming increasingly meaningless is the question of, well, how many hours did you work?

I would struggle to answer that.

I don't know what working is these days.

And I think, you know, some of my most productive moments are when I'm going for a walk or in the bathroom and sitting in front of a computer.

I could be incredibly unproductive in front of a computer.

And, you know, if someone said, well, how many, how many, how much time did you spend actually were actually working on that thing?

I don't even know if I could answer that if I was being watched, if they can answer that, you know, it's just, how do I know when I'm actually working on that and not.

Speaker C:

When your subconscious Is working.

Speaker B:

Right, exactly.

And I think it's not like you can measure someone, they're, are they on the production line banging the thing?

Are they doing that?

Oh, they've gone to the toilet, stop the clock.

Right.

You can't do that with these kinds of, this kind of work.

So I think, you know, that question, well, how many hours are we working?

That's hard to answer.

So I think increasingly, you know, all of this is starting to.

The whole kind of concept of measuring productivity is just becoming more and more challenging.

Speaker C:

I wonder what the stats would do if you just measured per head rather than per hour worked.

Speaker B:

I mean it would be meaningless because, you know, what do you do about part time employees?

You know, what do you, what do you do about people who only work a day, a week?

What about people who are on, you know, it's just like that's the whole point at the beginning.

I mean originally that's what people did do and then we started trying to measure.

Well actually what we care about is productivity per hour.

So that's a bit easier.

We can ask people how many hours they worked, you know, and now that's increasingly how much, how much time does, do people spend in the office actually working is my question.

We don't know and they don't even know.

Is having a chat with someone work?

Is a meeting work?

You know, is typing an email work?

Is it working when you're, when you're playing football after work with your colleagues.

Speaker A:

And team building, brilliant idea and yeah, establish a network and so on.

Speaker B:

Yeah, yeah.

In some regards.

I sort of think sometimes all of the, you know, the real value added that I might put in a project probably can be boiled down to about 10 minutes of insight and the rest.

Speaker C:

Of it's down the pub.

Building camaraderie, that's sort of thing building.

Speaker A:

Well actually just I think, I think it sort of fits this.

Just yesterday, as you know, I teach a little bit of media production and I said to him in my class, who are really, really good and very hard working this class of 16, 17 year olds and said, look, you're working too hard, you are good workers, you're very studious.

But that's not how this stuff works that we're talking about because they're doing a marketing campaign, right.

And I showed them some classic adverts like the Sony Bravia bouncing balls.

Oh yeah, right.

And other things, right, like dogs bouncing up and down on a trampoline for John Lewis.

Etc, etc.

I said I think these are great adverts and I think they're doing a really good job of what they do, and I suspect they would have been, you know, added to the bottom line of the, you know, respective companies.

But I'm telling you now, these ideas, the concepts, did not happen with someone sitting at a computer working.

Speaker B:

Okay.

Speaker A:

That's not how this stuff works.

And I. I think that's kind of what we're talking about.

Speaker B:

Absolutely.

Yeah.

And increasingly, I think more and more work inverted commas looks like that.

So, yeah, good for me.

Speaker A:

Brilliant.

All right.

Okay.

I think we're done.

Speaker B:

Yeah, why not?

Speaker A:

All right, so let's wrap up there.

Speaker C:

We've got to clock out now.

Speaker B:

Yeah, that's it.

We've.

We've done.

We've done what, 45 minutes.

Speaker C:

Wouldn't want to get in trouble with the unions.

Speaker A:

Exactly.

Speaker B:

Exactly.

Speaker A:

Exactly.

Okay, so you've been listening to the Cognitive Engineering Podcast brought to you by Aleph Insights and produced by me, Fraser McGruer.

If you haven't already, please like and subscribe.

We aim to release an episode every week or two.

If there are any topics you'd like us to cover, please email us podcast@alephinsights.com thanks, as always, for listening.

Until next time.

Goodbye,.

Speaker B:

Sam.

Show artwork for Cognitive Engineering

About the Podcast

Cognitive Engineering
Welcome to the Cognitive Engineering podcast.
Welcome to the Cognitive Engineering podcast. Occasionally coherent musings of Aleph Insights. We hope you like listening to them as much as we like recording them.

About your host

Profile picture for Fraser McGruer

Fraser McGruer