
J.A. Fdez. Madrigal
@_jafma
Followers
298
Following
39K
Media
569
Statuses
5K
Improbable mezcla de profesor, investigador, programador, escritor y otras cosas | Unlikely mixture of professor, researcher, programmer, author & others.
Joined October 2014
And we have definitely proved that scaling the computational power does not lead to intelligence.
The only positive thing I can say about generative AI is that we've learned a lot from its failures. For example, we now have even stronger evidence that human intelligence cannot be reduced to language processing.
0
0
1
Pero usar la IA en educación es muy güeno y fomenta el pensamiento crítico, y el cítrico también 😡.
This is the central problem with higher education in the age of AI. We can't require students to do take-home writing assignments (e.g. term papers) any more, because most will cheat and have ChatGPT or Claude or Grok do the writing. But we can't teach critical thinking,.
0
1
2
The best reason I've found so far for not using LLMs to assist writing (particularly in education):.
Yes. Writing is not a second thing that happens after thinking. The act of writing is an act of thinking. Writing *is* thinking. Students, academics, and anyone else who outsources their writing to LLMs will find their screens full of words and their minds emptied of thought.
0
0
0
This happens in any system of some complexity for which we know the outcome and little more. The processes that can produce that outcome can be so different and so many that in most cases it is impossible to recover the one that did the work; moreover: how can you know which was?.
Yep exactly this: Learning next sequence prediction doesn't recover the nature of the process that generated that sequence.
0
0
1