The Computer Thread

I use Otter.ai because it recognises and tags voices. But there’s no AI in there, it’s all just brute force learning. NLM queries are quite good but that’s not AI either.

I can see why they’re promoting you, this is the core of the issue and the reason any decently intelligent person has nothing to fear from AI, it’s no more threatening than a calculator.

1 Like

100% you still need a trained monkey to operate the tool, and if you’re not using all the tools at your disposal you’re a little bit crazy. Just don’t put IP or sensitive information into the “free” ai platform and you’re mostly good, dependant on how well as a monkey you have been trained.

[homer monkey clapping gif goes here]

We have a guy trying to sell us an AI platform at the moment for helping with bid responses. I can genuinely see some merit in it, you incorporate your win themes and client evaluation criteria into each specific bid within the platform and then plug in the RFP questions. It comes up with responses based on your past bid responses (you use your own best practice content as a library). It also connects to a creative ai + internet ai.

It’s fucking expensive though, the cost of an extra person - who can do more than one very specific task in comparison.

I’ve been trying to do the same thing, to ingest our past client maturity evaluations whcih include grading and descriptions of issues, and our recommendations. The purpose is to be able to generate a range of recommendations for a given issue based on commonality of prior recommendations. It simply isn’t reliable enough, the effort needed to weed out hallucinations and its determination to interpret our words in a subject where specific language is critical means I’d be better hiring another consultant than pursuing this.

But you have given me a lightbulb moment. I can codify the language used to describe the issues we identify and have it generate pick-lists for language consistent (and therefore analysable) assessment content that we can use to show clients how their issues compare in terms of commonality and frequency. Speeds up analysis documentation and provides an additonal normalised perspective. (Everyone struggles with this)

2 Likes

One of my queries was around their implementation - do they do any databasing or categorising of the library so that if we are doing pure project management vs commercial/transaction work, the tool doesn’t go to the wrong library to get content as the two are very different skill sets with massively different approaches, risks and challenges.

They didn’t which I found crazy. Not being able to select a few parameters to ensure you’re pulling the content from the right places seems like a huge oversight on their part. They were very much relying on the system using the right key words to find the right content. Given the words project and management feature heavily in both libraries I was not entirely convinced it’d perform that well.

It could generate content using a Bogan Australian language model though!

Most of the ‘AI’ businesses are so enamoured by their cleverness that they forget the most basic facts about reality, these models (and any comparative/interrogative model) rely heavily on high quality, consistent and unambiguous data to work properly.

2 Likes

And they can’t draw hands.

2 Likes

You can see how ai generated content is killing meaningful Google searches. I have to type in Reddit or other parameters on the end to try to get human input rather than total guff. I can’t believe they have let their golden goose platform become so degraded and useless

You and daveee are looking at this from your point of view and your work space.

I’m looking at it from my work space. The automation systems already exist to more or less fully automate bulk warehouse systems. I’ve worked in spaces with the different aspects, just not all in one place yet. There’s probably somewhere in the world with automated pick/packing as well.

Add in AI to understand and adapt to the scenarios that currently require human intervention and that’s … 30+, jobs where I am atm.

Looking to the future, mainly from an FMCG perspective, your customer just enters their needs into your system and that’s sales and inventory planning jobs gone. Move on from that, BOM and raw materials doesn’t need people.

Will it happen in my working life? Probably not. Will it chip away? Yes, it will.

And we contextualise that with the discourse that individual worth is validated by participation in the work place. Whilst we continue to reduce the need for labour we need to address how worth in society is addressed.

How many train drivers are driving these new metro trains? How many less ticket collectors are there on trains now than in the 1980s?

How many station masters compared to the 80s and 90s?

Bus conductors?

How many less 15 year olds in McDonalds now?

But jobs are important. Just ask any politician right? But only in the Gallilee Basin …

Technology has been ‘chipping away’ at jobs since humans started using tools.

Jobs will disappear, new jobs will take their place, there’s always someone fearmongering about the impact of the latest wave of technology but living standards continue to grow and unemployment rates vary within a very narrow band regardless of socioeconomic status (Minority segments with structural challenges aside). Mainframes, PC’s, robotics and the internet have all failed to collapse the labour market, AI will fail in the same way.

Of course there are always individual stories of personal impact, people too old to retrain or lacking the motivation, but the sky has yet to fall in, society adapts and trains the workforce for the new reality.

2 Likes

I want to see what the new reality is!

Historically, the fear mongering about being replaced by machines has been in an effort to supress wages. Better not ask for an increase in minimum wage.

1 Like

It’s also based on the fact that people just really really don’t like change and it scares them. For people that have been in the workforce for 40 years, it’s not exactly easy to find out they’re going to potentially be replaced by a computer system that’ll do their job faster and cheaper. Yes that system has an entire new workforce around them, but it still means the older generation is suddenly having to shift work etc. with low chances of either getting hired or even getting retrained

The brutal fact is that that’s how capitalism works, while it may be ‘unfair’ or at least callous at an individual level, the eveolution and improvement of workplace practices is what drives the efficiencies which raise everyones standard of living*

*yes, I know about the growing gap between efficiency increase and wages, I know about the market dominance in specific sectors that have allowed real wages and living standards to stagnate and even regress, but that’s a recent phenomenon, the model has worked well in the past.

100% agree with you. The only problem is that with medical advances, old people are simply living longer, so the period between end of work and death is growing, which means the Government is forced to spend more resources on a generation that isn’t productive (as blunt and cruel as that sounds). There needs to be huge investments from the governments to upskill older generations to keep them in the workforce, as well as increasing the retirement age. Japan is already having the issue with an aging population on trying to figure out where the money is going to come from to support them.

I am not entirely convinced of this.

Unless they are a waste of oxygen, there is no reason they cannot specialise in their field of work - people still need to enable whatever function is being replaced by a computer. Someone needs to tell that computer how to operate, what it needs to consider. Then it actually needs implementing too.

Jobs still exist, how they can apply their specific skillset to a changing industry is generally how the world evolves.

Training your run of the mill 60 year old on using computers isn’t an easy process though.

So change was forced on the Agrarian Peasant system as all the peasants left the fields to work in the new factories and become the working class.

As efficiencies in Industry increased and decreased it’s need for workers those workers moved into new service rolls, like transport and telecommunications and bin collection and mail delivery. Council park maintainers. Tip attendants and plant operators. The increased scale and size of transport people meant more people into road building/maintenance.

At the end of the workforce I’m in, I still see people coming in now that are not competent with computers. That’s basically 2 generations after the home computer revolution.

They’re not running systems, they’re not tech savvy. There’s a large percentage of our population that can’t handle, mentally, sitting at a computer all day. Is there going to be service roles or labour roles for them when automation ramps up?

Maybe it’s time to address the capitalist system? Is it serving the people the way they need to be?

There’s other ways of conducting society. Is capitalism still the least worst system?

The thing is, young kids can’t use computers for shit past the surface level anymore either. Since everything is automatically installed apps, they’re just as lost one you go one level down.

4 Likes

This is a good read into Chile’s Project Cybersyn - a project set up in the 70’s to build a bottom up planned economy. Super intersting stuff. Modern computing and AI can definitely be used for the common good and not just for stealing jobs.

Project Cybersyn: Chile’s Radical Experiment in Cybernetic Socialism | The MIT Press Reader

Edit: Paul Cockshott’s book “Towards A New Socialism” is great, and was inspired by Chile’s Project Cybersyn. Keep in mind that this was written in 1993, obviously the level of computational power we have now is obscene in comparison.
Link

1 Like