Are you being served?

 In news

The sleeves will ride up with wear

Those of you reading this article having celebrated a significant birthday beyond 21 with 30 not counting for anything (it never did, get over yourself!) might recall a certain tech development methodology deployed by Mr Gates and Microsoft all those years ago. It involved sending to market, software that had ‘bugs’ knowing that mass usage would detect faults quickly and rapid repairs by ‘patches’ would restore service. Annoying to some, irrelevant to many of us because those were the days of ‘dark arts’ IT departments with most tech being encountered by individuals within a work environment where help was on hand and telephone enquiries were permitted and answered. And it’s a fact – the real world of mass usage is the quickest way to develop this tech for the benefit of us all. Just perhaps not always the wisest or fairest – in some cases.

I can see it all now…

What you might not know is this self-same development strategy is alive and well, as evidenced in The Sunday Times (February 16, 2025, D. Fortson) and being deployed by Sam Altman’s Open AI. Described as:

“…an iterative approach: pushing out flawed products so that the public can use them and provide feedback as a way to fast-track improvement.”

It is reported this process is being used with the AI products packages Operator and Deep Research.

Just an observation but maybe this approach has the added benefit of being more difficult to copy and IP protect as it shape-shifts its way forwards – incrementally – quite possibly all by itself. Can the patent office keep up we might reasonably ask.

I guess we all just pray this approach to development never finds its way to the defence or medical industries. Or maybe it has already? Ooh…err missus as Frankie might have said. Carry on nurse, plug me in for my software back-up.

And good to know that ‘Billy 4 Jets’, the bug-meister, was ahead of his time. Of course, now, if you believe conspiracy theorists, he’s asking the world to eat his bugs, crickets mostly by all accounts, having purchased an estimated 270,000 acres of farmland in the USA.

You’ve all done very well…

Boffins in AI are saying the current surge in activity, development and interest is around Agents. In short, that’s an AI tool that, like a browser, goes onto the internet to carry out tasks for the user whether that’s serious research or booking a rugby ticket. Sometimes called “super assistants”. And these tools are truly amazing – even now. One commentator described it like having your own PhD research assistant only quicker, much quicker. AI’s capabilities are accelerating at speed. We should perhaps ask what controls are there over this runaway mustang? Before it bucks.

Take comfort, or not, from Altman when he says of AI development:

“…I think there will be a lot of good and a lot of bad. My hope and my belief is that there’ll be orders of magnitude more good than bad…”

Phew – that’s all right then.

“…but there will be real bad…”

Which is it Sam because we’re alarmed that you simply don’t know. And will it be democratic or authoritarian?

Open AI is being used by more than 300 million people a week; my hunch is they’re not all saints.

And I am unanimous in that…

By way of timely example Sky News recently reported that AfD in Germany was benefitting from the use of automated AI-assisted social media messages and even the AI creation of

“…deepfake videos of apparent ‘witnesses’ or ‘whistle-blowers’ fabricating stories about prominent politicians.”  

Disinformation ‘flooding’ social media to influence political campaigns is perhaps just the starting point for bad actors in AI. AI generated social influencers now exist that can say radical things that break national anti-hate/discrimination laws largely free from the threat of the rule of law. You can’t throw them in jail because their ‘puppeteers’ are multi-jurisdictional and pan-global, difficult to locate and tricky to catch.

Has anyone told Mr Rumbold.

Going up.

 

Murray Fairclough

Marketing Director

OPUS Underwriting Limited 

+44 (0) 780 145 9940

underwriting@opusunderwriting.com

Recent Posts