[LINK] OpenAI’s o3 Model: Breakthrough or Breakdown?

[David]

On Monday, 28 April 2025 14:57:33 AEST Scott Howard wrote:

> For a simple case like the one you've mentioned they probably just do it in their head, but in a more generic sense the easy answer is to outsource that calculation to something designed explicitly to do that type of action - such as a calculator.  In an agentic AI world, the role of the AI isn't to add those 2 numbers together, but to outsource it to an agent (such as a calculator, via an API) that will do it for it.\

But isn't that just moving the problem out of sight?  The original issue was that an AI system would not recognise the difference between the hard logical proposition 2+2=4 and the softer existential (empirical) proposition "2 plus 2 usually makes 4" (using these as illustrative rather than literal examples).  Even an agenic AI system with access to the appropriate resource(s) still has to initially recognise two distinct cases.

Then there's the issue of ultimate legal responsibility, which has to be carried by a warm human being.  If I were an Engineer designing a structure or a Surgeon planning a procedure where safety of life was critical, I'd want to make damned sure any AI system, agenic or not, could be relied upon to the point where human lives and my professional reputation were not in danger.  So I imagine there would soon come a point where the human involved might as well do the whole job themselves.

I think AI systems are very important and extremely useful in many situations, but we need to know their constraints and how to use them.  

_David Lochrin_

-------
> This seems to me to demonstrate a fundamental issue with AI machines: they
> > are still very large correlation processors but have not been developed far
> > enough to distinguish "empirical correlations" and logical rules.  Thus "2
> > plus 2 usually makes 4" is an empirical correlation, but "2+2=4" in an
> > appropriate mathematical context is a logical rule.  (And we understand "He
> > added 2 and 2 and got 5" isn't an appropriate context!)
> 
> 
> This is one of the many areas where the idea of 'agentic AI' comes in.  If
> you ask a human to add two numbers, what would they do?  
> 
> Mix this with the newer reasoning models which are much better at working
> out the best path to come to an answer, and the types of answers you get
> from AI systems now days for this type of question is significantly better
> than it was only a few months ago.  Instead of simply looking at "2 plus 2"
> and trying to guess what comes next, the recent models are able to look at
> that statement, determine it's a calculation, decide that the best way to
> solve such a statement is by using a calculator, and then call a calculator
> agent to actually do the work and get the answer.  If the calculation was
> even more complex, then they might instead decide that the best option is
> to write python or R code to solve it, run that code, and then return the
> answer.
> 
>   Scott
> _______________________________________________
> Link mailing list
> Link at anu.edu.au
> https://mailman.anu.edu.au/mailman/listinfo/link
>