Show newer

@bortzmeyer
Incroyable - j'ai demandé la même chose, écrite en "R" ... pas de problème. Et il (elle ?) sait que "Il peut ne pas être pratique de lister tous les nombres réels dans une plage donnée." C'est ça.

Je demande à #ChatGPT un programme qui énumère les nombres jusqu'à N. Il produit un programme pour des nombres entiers. Espérant le confusionner, je lui demande un programme qui marche avec les nombres réels. Il me fait une variante qui énumère avec un pas (arbitraire) de 0,1.
Dois-je le considérer comme très bête ou comme très intelligent d'avoir interprété de manière créative un cahier des charges absurde ?

@boris_steipe @amills And the initial genesis thought seems like the most important one. Whatever comes to #mind, we try to make it true, with the actions and words we choose after that. An important social skill is to be #consistent in our beliefs/values/personality with ourselves & our peers. To let a mindless machine initiate that self-storytelling is giving up the most powerful moment of #FreeWill we have. Don't let students cede that #agency to an AI #agent.

I’m concerned about the assumption that if an AI can do something it’s not worth learning to do unaided. It can write a pseudo personal statement but students need to learn to write an authentic one.

Show thread

@amills

That's a point worth thinking about. I'm sure we already agree that _any_ learning is already worth doing, for its own sake. And that being able to introspect, and put such introspection into words for a personal statement is a cultural skill that touches on many dimensions. And that an AI synthesized personal statement is an oxymoron.

Let's think about two different skills however: the skill to write a personal statement unaided, and the skill to take a synthesized draft, spot and understand its flaws, and make it so much better. Do these skills overlap? Completely? Causally?

@jhackett @bookstodon

John - this is fascinating! But you need to realize that those books do not actually exist (except for the last one - and, well, The Yugoslav Experiment is a chapter in Singleton's book).

I have written about those "Schrödinger Facts" in my latest newsletter for the Sentient Syllabus Project – and also why that information is not useless.

Have a look: sentientsyllabus.substack.com/

@hfueller @robertlepenies
Doch - denn wie sieht die Alternative aus? Entweder wir lernen die AI als Trittbrett zu nutzen - oder sie wird zur Konkurrenz. Und diesen Konkurrenzvergleich verlieren wir ganz schnell: wir werden nämlich nicht klüger. Und wo ist die Zukunft für meine Studenten, wenn die mit einem Algorithmus konkurrieren, der 24/7 ohne urlaub, Sozialabgaben und Mutterschutz schafft? Wir müssen besser sein. Und das geht nur wenn wir uns auf die Schultern der AI stellen - egal wie gut die ist.

Energiebilanz? Ich weiss nicht ... das klingt alles nach recht viel, ist aber im größeren Zusammenhang fast nichts. Es werden ja dabei auch wieder Energien eingespart, bspw. wenn wir weniger Google Suchen brauchen, weniger Konzepte ausdrucken ... ich kenne die Arbeiten, bin nicht überzeugt. Die Frage nach der Menschenwürde - d.h. wo für unsere Kreativität und Können noch Bedarf besteht - das ist doch viel, viel wichtiger.

Aber bewegen kann man da nur was wenn man win-win Situationen schafft.

@amills

Great work Anna, very helpful. I added two links to resources we have at The Sentient Syllabus Project (sentientsyllabus.org)

@cysec Im Sentient Syllabus Project versuchen wir darauf robuste, praktische Antworten zu finden. Schauen Si mal rein - sentientsyllabus.substack.com und sentientsyllabus.org - und teilen Sie's wenn Sie jemanden kenne dem das helfen könnte.

@hfueller @robertlepenies

Im Sentient Syllabus Project arbeiten wir an robusten, praktischen Ressourcen für die AI Era. Scheuen Sie mal rein - sentientsyllabus.substack.com und sentientsyllabus.org - und teilen Sie's wenn Sie jemand kennen dem das helfen könnte.

@ColinTheMathmo

There are other great uses too. Playing with knowledge, refining your grammar, having a personal tutor ... have a look at the Sentient Syllabus Project if you are in education. sentientsyllabus.substack.com and sentientsyllabus.org

@marcidale Yes, good read. But check out the Sentient Syllabus Project for more in-depth academia and AI resources to share with colleagues: sentientsyllabus.substack.com and sentientsyllabus.org

ChatGPT's Achilles Heel

I just posted a discussion of ChatGPT's major weakness for the Sentient Syllabus Project: sentientsyllabus.substack.com/

It falls prey to its "Schrödinger Facts".

In a nutshell: ask the AI for sources - and then check whether they actually exist. They usually don't. Requests for specific, provable sources turn out to be ChatGPT’s achilles heel.

That's a big deal: (a) we can (and should) always require students to attribute ideas anyway, and chasing down the source of a ChatGPT claim may be more work than working from an actual source; (b) ChatGPTs information is not useless though - and can be very valuable to point to the right direction; (c) while we become masters in attribution, the collaboration benefits in other ways, and the end result is improved.

Find our resources on academia and the new AI at sentientsyllabus.substack.com and on sentientsyllabus.org

... and share this information to whoever might find it useful.

Three principles for the academy in the era of digital thought:

1: An AI cannot pass a course.

2: AI contributions must be attributed and true.

3: AI use should be open and documented.

I explain why in our latest post for the Sentient Syllabus Project. Have a look at sentientsyllabus.substack.com

@debasisg@mstdn.social
Well ... don't forget its data cuts off in 2021.

The Sentient Syllabus Project
Charting a course for academia in the era of digital thought.
--

I've been really busy over the last days rolling this out: the large-scale availability of literate AI has imploded the academy as we know it. We need to re-orient. How? I founded The Sentient Syllabus Project to share concrete, practicable advice. Check it out at sentientsyllabus.org and at sentientsyllabus.substack.com/

Most importantly: this includes text samples how to start re-thinking and adapting our winter-term syllabi by next week. This is needed _now_. Drop in, have a look, and if you know someone who might benefit please share.

"A beginning is the time for taking the most delicate care that the balances are correct."

(Frank Herbert, Dune - 1965)
--

I was a teenager when I first read Dune, waiting in the queues of the Stubai glacier ski-lifts, in Austria. At the top of the glacier, a short hike would bring you to the saddle and the view from there soars over an untouched snowy plain, beckoning, sparkling, dipping into the Windach Valley high above Sölden, and the view leaps across the slope and continues all the way to the blue peaks of Italy's Dolomites that make the horizon. I stuck my skis in a drift, sat on a heap of scree, chewed on an apple and read.

--

So this is a beginning. Hello, Mastodon. Are the balances correct? The thing is, one cannot know that, at the beginning. Such knowing is for the end. Until then, there is a path. Let's see where it leads to.

In lieu of an just tags: and , programming and teaching , , and and ... but also and more recent published work in , and - jointly with @yi_chen .

Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.