Re: AI slop "contributions" to IETF working groups

[Phillip Hallam-Baker <[e-mail]>]

On Tue, Feb 10, 2026 at 9:55 PM George Michaelson <[e-mail]> wrote:

> Although I agree with you, inventing new terms-of-art like SI is not
> helpful. I blame the AI industry for using english language concepts like
> "hallucination" enough as it is. The superficiality is true, but the field
> has 50+ years of AI as a key concept. It was misapplied when it was shakey
> the robot, it's misapplied now.
>

My intention was to make a point, not establish a new term of art. Not that
I see a problem in 'superficial intelligence' as a subdivision of the field.

To be serious for a moment, back when I was in the AI lab, I asked Patrick
Winston if he thought the goal of the lab was replicating human
intelligence or understanding what it is. His response was 'a bit of both I
think'. The Web represented a third approach, providing mental prostheses,
tools that extend and expand intelligence.

What superficial intelligence technologies are showing us is that the bar
to appearing intelligent is rather lower than we assumed. LLMs have no
reasoning capabilities, they are working purely by interpolation. So maybe
we need to rethink our model of animal intelligence at the very least.


> So thats Yes, but also No. please, don't do this. Its not SI. Its AI and
> LLM and GPT because thats what people "out there" understand.
>

Currently we have several hundred billion dollars being dumped into AI with
no visible business model to provide a return on that investment. Many of
the carnival barkers who are selling these dubious business propositions
are the same people who not so long ago were selling people NFTs.

I think that introducing the term 'Superficial Intelligence' into the
discussion would allow a rather less exuberant mode of discussion.