I asked it 'what makes good climate legislation'. It answered, in seconds, with machine generated text:
1/10
Ok, I thought, amazing! Though a bit light touch and generic. I pressed 'try again'. Whoa sweet baby! It gave me multi-sectoral approach! affected commuties! Science based! Coordination! Transparency!
2/10
How much can we push it for actual policy measures?! Let's inquire about incentives for renewable energy.
Ready?! I'm not sure you are:
3/10
Ok. At this point I'm beyond words. I ask it to tell me about successful renewable policies in Asia and how much investment those facilitated.
It answers:
4/10
I WANT IT TO COMPARE STUFF NOW. IT DOES.
5/10
But I'm tempted to press try again for that one too. It delivers:
6/10
Hmmm can I get it to write a law? @bryworthington has been dreaming of this:
ARE YOU SITTING? Do you have triple shot of whisky handy? It wrote a law. 🥃🥃🥃 Le'chaim.
7/10
To summarise: the last hours, the internet has exploded with people throwing crazy questions at #GPTchat (don't even get it started on cats..). I'm beyond convinced that it is about to revolutionise the world of climate law and policy research and policymaking.
8/10
It'll be amazing, it'll be amusing, it'll be mega-efficient, & it will come with a whole bag of horrors to consider. We'll have to test, verify, anti-bias and all the things - and it'll still be scary. The utopia-dystopia divide isn't a simplistic crossroad. We'll have both.
9/10
Our responsibility (all of us, but users of large language models, & we at @climatepolradar in particular) is to actively play a role in making sure that this technology is used for good. We'll be thinking a lot about how we can do that & welcome thoughts & collaborations.
10/10
11/10 A few months ago, @L_P_Jones asked #GPT3 (the older sibling of #GPchat) the big question, and I'm not sure if he was pleased with the answer, or whether he preferred early retirement and cocktails on the beach 🍹🏖️