[ExI] Mono or Poly?

[BillK]

On Mon, 24 Feb 2025 at 08:21, Rafal Smigrodzki via extropy-chat <
extropy-chat at lists.extropy.org> wrote:
>
<snip>
>
> So, I don't know. I am genuinely perplexed. Should we throw all our eggs
in one basket, have a trillion dollar computing Manhattan Project to build
the One AI as soon as possible before the Chinese communists kill us all,
or should we rain money on a dozen startups and let a hundred flowers
bloom?
>
> If you are concerned with frightful omens, the Manhattan Project gave us
the bomb and the hundred flowers were killed by Mao Zedong in the
subsequent Anti-Rightist campaign, so maybe we are screwed either way.
>
> For a long time I was mono, now I am dipping my toes into poly. Poly
feels good for now - but is it good in the long term?
>
> Are you guys mono or poly?
> --
> Rafal Smigrodzki, MD-PhD
> Schuyler Biotech PLLC
> _______________________________________________



OK, so I asked a couple of AIs for an opinion.
They produced reports of the pros and cons and in effect agreed with Rafal
that it was impossible to decide which option was best.
Just as well we don't have to make that decision, as the race to AGI is now
unstoppable.
BillK


OpenAI o3-mini
Conclusion

The article ultimately leaves us with a profound uncertainty:

   - *Should society channel immense resources into building a single,
   controllable AGI (a Manhattan Project of sorts) with the hope that careful,
   deliberate progression ensures safety?*
   - *Or is it better to support a diverse ecosystem of AGI developers,
   trusting that inter-AI competition and cooperation might self-regulate the
   overall system, even if it requires accepting a greater degree of
   short-term risk?*

The answer is not clear-cut. Both scenarios have significant trade-offs:

   - The *mono* approach has the benefit of deliberation and the luxury to
   think—but it's a high-risk strategy if its path diverges from intent.
   - The *poly* approach offers a form of natural checks and balances
   through competition, yet it may inadvertently incentivize risky behavior
   due to the urgency of outpacing competitors.

As the article notes, whether we “throw all our eggs in one basket” or
“rain money on a dozen startups” remains an open question, with significant
historical analogies that show both centralized power and multiplicity can
lead to dramatic consequences. At present, the dilemma is one without an
obviously optimal solution, and much hinges on further research,
technological safeguards, and possibly the creation of frameworks that
might combine the merits of both approaches

------------------------


<https://poe.com/DeepSeek-R1-FW>
DeepSeek-R1-FW

*Conclusion*

The choice hinges on risk tolerance and technical feasibility. A
monocentric AGI offers a clearer alignment pathway but risks catastrophic
failure. A polycentric system spreads risk but introduces chaotic
competition. Neither is inherently safe, but the polycentric model’s
redundancy might offer a pragmatic middle ground if combined with robust
governance. Ultimately, the decision must balance urgency (e.g.,
geopolitical competition) with safeguards, recognizing that AGI development
is as much a political and ethical challenge as a technical one.

-----------------------------------
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20250224/ce702955/attachment.htm>