Les Orchard
Whew, this ChatGPT shit is just a great big security exploit against human theory of mind innit?
I'm being continually astonished by folks attributing confidence, intent, and lying to this thing that's just a clever algorithm recombining source material.
It's not a confident actor lying with intent, it's a system serving up confident lies sourced from human actors
Adrian Cochrane
@crashglasshouses @lmorchard I don't recall that scene?
Inken Paper 🇵🇸
@lmorchard there's a scene in Young Frankenstein about this :blobcat3c:
Les Orchard
When I was a little kid, I attached a blob of playdoh to an electric motor and let it flop around on a table.
For a good 10 minutes, I was spooked and convinced I'd made something alive. It moved rather randomly and even seemed to lunge at me occasionally.
But, you know, it was not in fact alive. Any life it apparently had was assigned & extrapolated by my own mind. Handy for quickly reacting to things that might be coming for you through the trees, but not exactly an accurate instrument of assessment.
Daniel Taylor
@lmorchard it's not even lying, as it would have to have knowledge of truth to do so.
It's blathering, it just emits the words that seem to match the current situation (according to its algorithm) without care or concern to values of truth or falsehood.
It would make a very effective political campaigner
Adrian Cochrane
@RandomDamage @lmorchard Or Silicon Valley entrepreneur! Who thought they'd be the ones to automate themselves auto of the job?
０７６萌ＳＮＳ is a social network, courtesy of ０７６. It runs on GNU social, version 2.0.2-beta0, available under the GNU Affero General Public License.
All ０７６萌ＳＮＳ content and data are available under the Creative Commons Attribution 3.0 license.