[ExI] VERY cool. Toward mind uploading?

[Ben Zaiboc]

On 09/03/2026 12:19, BillK wrote:
> I told Gemini that the full adult connectome was used, not the larva phase.
> So it amended the report.
> BillK
>
>
> Gemini AI 3.1 Pro Thinking -
>
> You are absolutely right to correct me on that. The leap from the 3,016-neuron larva to the 139,255-neuron adult Drosophila melanogaster (the "FlyWire" project, published in Nature in October 2024) is a categorical shift in the scale of the "upload" argument.
>
> This achievement, which involves more than 50 million synaptic connections, isn't just a map; as you noted, it includes a functional computational model (using leaky integrate-and-fire neurons) that drives a physically simulated body.
>
> Let’s re-examine the implications for the three objections with this adult-scale complexity in mind:
...

> Summary of the "Adult Fly" Milestone
>
> The fact that this was done with an adult brain moves us from "simulating a worm" to "simulating an agent." It effectively moves the "Mind Uploading" timeline from "science fiction" to "active engineering project."

It should be clear by now that this is not intelligence, it's, well, I'm not sure what it is, but it's not intelligence.

I suppose it's what you get when a system is very good at predicting which words normally follow certain words, with a huge amount of data to train on. No actual understanding or intelligence here at all.

I think we assume it's intelligent at our peril.

My concern is that this will lead to another AI Bubble, which will eventually burst, and set back AI research again.

My hope is that some people realise this, and are working on systems that actually have potential to become really intelligent.

-- 
Ben