AI - Is the end of the world (as we know it) nigh?

RE: AI - Is the end of the world (as we know it) nigh?

Arlo Guthrie — Sat, 21 Mar 2026 17:12:18 GMT

[quote userid="3169" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh/248559#248559"]and it's very worrying if you go down that wormhole. [/quote]

But wouldn't it be a useful thing if there was a more objective way to better differentiate between the good studies and ...

[quote userid="3169" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh/248559#248559"]There are some awful studies out there, low numbers, inadequate power, spurious conclusions etc. So many more done in referral populations that mean the results are not relevant to general practice populations. [/quote]

Your last point in particular is really interesting, because my thinking is that ai could calculate the degree to which it is or isn't possible to infer. So it would highlight when it is not possible to infer that a finding in a particular group could be applicable to a broader group.

I think if nothing else it could be an interesting intellectual exercise, and I'm going to try something fairly rudimentary and see whether it leads anywhere (or not) ...

RE: AI - Is the end of the world (as we know it) nigh?

Michael Woodhouse — Sat, 21 Mar 2026 00:51:36 GMT

[quote userid="2100" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh/248549#248549"]I’m imagining a tool which could objectively tell a busy practitioner whether a study is robust, or, er, not.[/quote]

There are a few studies in vet med with large numbers of animals, good results, run well.

THE study, often cited, that a TPLO is better than a lateral suture has 80 dogs in it - such low numbers, such low power. You are legally allowed 30 kids in a primary school class. There are some awful studies out there, low numbers, inadequate power, spurious conclusions etc. So many more done in referral populations that mean the results are not relevant to general practice populations. We seldom have a negative control group and simply ignore regression to the mean.

As a new grad I would have given an answer I was 100% certain about how to treat endometritis in cows. I did my masters (DBR) and looked at a lot of research in great detail and I'm now not very sure. Seemingly there are many ways to treat the same condition, there's even published papers that some treatments are effective, when subsequent testing has proven they are not. Remembering that we are "happy" at 0,05 so 1 in 20 results due to chance. Or bias. Or lies.

We "know" so very little, EBVM is a borderline sham and it's very worrying if you go down that wormhole.

RE: AI - Is the end of the world (as we know it) nigh?

Cinzia Gandini — Fri, 20 Mar 2026 12:05:01 GMT

[quote userid="2100" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh"]percentage of knowledge-based jobs[/quote]

It was just reported in the news that the employees at InvestCloud were software engineers and IT specialists who developed and customized financial software platforms. They were laid off as part of a restructuring in which some of their work is being replaced by AI systems. The company is shifting toward a centralised, AI-driven platform, reducing the need for local development teams.At the same time, InvestCloud says it is still hiring, but mainly in New York, meaning affected workers would need to relocate to continue working for the company. The employees at InvestCloud were software engineers and IT specialists who developed and customized financial software platforms. They were laid off as part of a restructuring in which some of their work is being replaced by AI systems. The company is shifting toward a centralised, AI-driven platform, reducing the need for local development teams. At the same time, InvestCloud says it is still hiring, but mainly in New York, meaning affected workers would need to relocate to keep working for the company.

RE: AI - Is the end of the world (as we know it) nigh?

Arlo Guthrie — Fri, 20 Mar 2026 09:26:18 GMT

[quote userid="8991" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh/248490#248490"]For my sins, I do some peer reviewing[/quote]

David Mills Was thinking about exactly this this morning.

Problem: Many people (self included) are either not skilled enough or don’t have the time to critically appraise studies (or even more basic market research type surveys. They jump to the conclusion and take it as fact.

And of course the peer-review process is not infallible (by a long shot).

The Cite tool is supposes to give an indicator of the quality of a study, based on how often it is cited by others. But the citers themselves may not have checked the quality of the paper they cite. I’ve seen that happen.

What if we could create a system which reads studies, assesses them according to agreed criteria (no of participants, selection process, blinding, framing of question etc) and then produces both a strength indicator for the degree to which the study proves the hypothesis, AND what may or may not be possible to infer from it, and summarise that in an easily digestible form?

I’m imagining a tool which could objectively tell a busy practitioner whether a study is robust, or, er, not.

Would that be interesting?

RE: AI - Is the end of the world (as we know it) nigh?

Cinzia Gandini — Fri, 20 Mar 2026 05:32:35 GMT

[quote userid="3607" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh/248391#248391"] But what about dirty-hands work?[/quote]

I think they will disappear too, most have already gone now, the rest will be going soon

AI machinery refers to machines that use artificial intelligence to perform tasks that normally require human thinking, decision-making, or physical effort. Instead of just following simple programmed instructions, these machines can learn, adapt, and respond to their environment.

AI machinery is about smart machines replacing or assisting human work, especially in repetitive, dangerous, or highly precise tasks.

Forestry
Foresters cutting trees – 1 person using a machine, smart harvesters

Forestry uses drones with computer vision to
Detect diseased trees
Monitor forest health
Track illegal logging

Farming
People picking crops – machines do it
People pulling weeds – robots do it

Factories
Workers building things – robot arms

Warehouses
Workers picking items – robots

Shops
Cashiers – self-checkout

Customer service
Call centre workers – chatbots like ChatGPT

Driving (future)
Drivers – self-driving vehicles

Office work
Data entry workers – AI software

Cleaning
Cleaners – robot vacuums

and many others

RE: AI - Is the end of the world (as we know it) nigh?

Chris Milligan — Fri, 20 Mar 2026 04:39:02 GMT

[quote userid="3607" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh/248391#248391"]those for whom work is wielding a shovel, emptying a trawl-net and gutting the resulting fish, wiping the arse of your no-longer continent relative, comforting a grieving parent or reassuring an anxious patient, putting out fires, catching bad guys, rescuing errant hill-walkers or boaters (there is still life beyond video games!), schooling and educating (two very different things). And then there is art in its many forms. [/quote]

this is an argument I often see made about AI, but it isn't really true, it's just wishful thinking. I think one of the things that people struggle to get their heads around with AI, and particuarly AGI when it arrives, is that it will, ultimately be better at these things than we are, including artwork and film making and painting and plumbing. It's not that it will be able to replicate what the best of us can do it in each field, but rather it will fundamentally understand it better than we do. I think we'll see this first in the film industry - sooner or later it will be making movies from scratch (including the soundtrack, script, special effects and marketing) better than we can. that's what superintelligence means. Nick Bostrom and Max Tegmark write very elegantly about this in their books Superintelligence and Human 3.0 respectively.

RE: AI - Is the end of the world (as we know it) nigh?

Cinzia Gandini — Thu, 19 Mar 2026 18:35:42 GMT

[quote userid="2100" url="~/f/non-clinical-questions/31311/ai---is-the-end-of-the-world-as-we-know-it-nigh"]What do you think?[/quote]