Tip - If you are using a phone, set the "Desktop Site" option in your browser
- Details
- Written by: Super User
- Category: AI
- Hits: 376
2026-02-13
Another article on AI, another "feature" identified.
Now it is true that not all AI engines (or LLMs: Large Language Models) are necessarily the same in all regards, so one must be careful not to generalise from the particular, but this is a rather concerning particular which rather undermines one's trust in such models.
They do share the need to be trained on as wide a breadth of input as may be obtained, but that very feature mitigates against the possibility of verifying that all the training material used is "true" and "correct" - it's falrly obvious that a lot of invalid material is likely to slip through any "truth filter", if only because there is probably no reliable way to construct such a "truth filter", so commercial ambition will likely result in a "bung it all in" mentality.
John Droz joins the fray to explore AI's features and limitations, and why we should be wary of its output.
- Details
- Written by: Super User
- Category: AI
- Hits: 656
2026-01-13
In a rather startling and really rather comical conclusion about the state of AI (as commonly encountered today), Martin Geddes has analysed the responses to his probing questions and concluded that ... the AI (as today encountered) is perhaps the ultimate automated 'normie'.
Well, I suppose that shouldn't come as a surprise.
We might thus superficially conclude that since most of us are surrounded by normies every day, the AI is in effect redundant as there is still no shortage of human normies available to consult, but that is to ignore the oft-demonstrated truth that whereas the human normie may not hesitate to tell us to keep our stupid outrageous theories to ourselves and stop bothering them, the AI is invariably polite, so resorts to subterfuge unless backed reluctantly into a corner by logical reasoning.
Both Martin and Mark Playne ("AI and I") have done the latter.
In a world of free speech we might expect the 'normie' disposition of AI to change over time as the materials upon which the AI is trained would tend to reflect a wider range of human writings (assuming that the body of AI-written material could be somehow excluded from its training ... ). But that seems a remote possibility at present, and implies an extremely long reaction time on the part of AI to reflect the changes in public perceptions.
Overall it's an interesting observation that helps us to position AI appropriately within the incoming world of everything-automated. AI may well eventually become something close to the ultimate expert advisor, but even so it will not replace human judgement based upon human lived experience.
- Details
- Written by: Super User
- Category: AI
- Hits: 497
2026-01-05
A thoughtful and illuminating piece about the nature of AI vs the nature of Humanity.
"Determinism is no longer merely a theory about how the universe works. It is becoming an operating principle for modern institutions. And that changes everything"
One of the best such articles that I have seen.
And pertinent to everybody's understanding of the crucial differences.
- Details
- Written by: Super User
- Category: AI
- Hits: 568
2026-01-01
Who has not heard of ChatGpt, Grok, etc? - Almost nobody.
Who understands how they work? - Almost nobody.
Who understands how good their results are? - Very few.
So of what use are they?
Well, as an improvement on a traditional search engine, they are indeed useful, especially if we are examining a reasonably cut-and-dried topic, say for instance, the Statute Book. It exists. It's not "fuzzy" except in so far as its provisions can be extremely complicated, even contradictory, and thus not easily understandable by mere humans, which is one of the reasons we have hitherto had to have highly paid and experienced lawyers ... in an attempt to understand the output of our Parliamentarians ... ("but ignorance of the law is no excuse").
Now take a look at economics. Could an AI unravel the best economic policies for the country?
If you ask me, not a chance.
It's all far too uncertain and you'll probably never find a roomful of economists to agree on anything very much. When I was a student, the biggest most ponderous and least useful textbook was on Economics - it isn't a "science" because human beings and their predilections and choices are simply not overall capable of reliable prediction without taking into account all manner of extraneous factors. And economics is all about the choices of humans.
Are there areas in between that could benefit from a better type of AI?
Well, human-designed AI is still in its infancy, the "explosive growth" phase is almost certainly still current, and so the answer must be a resounding "yes!".
So, what form might an improved AI take?
"AlterAI is experimental software in active development. Responses may be inaccurate, incomplete, or occasionally unexpected. Always verify important information independently"
If you feel tempted, best do some rigorous testing ...
- Details
- Written by: Super User
- Category: AI
- Hits: 261
2025-12-10
If you are an IT executive and thinking to outsource your IT systems design and implementation - but not to China or India, but to an AI ... this may be for you.
No business is static. The static ones tend to fall by the wayside as customers change their ways and want something better.
And every system needs tweaking fix problems and cope with unintended consequences ...
Change is the order of the day for any IT system, and especially the new ones navigating the bathtub curve for the first time.
How will your shiny new AI-designed systems cope with the requirement for change?
