Browse Comments — Raw (as collected)

This is big news! I'm one of the few brave souls who has used Gemini CLI. It can be incredibly powerful but generally needs to be strictly harnessed by another frontier LLM. Started poking around in Agy. So far so good. There doesn't seem to be traditional token or usage limits. When I asked how much I've used it says it doesn't measure tokens the same way. Instead every convo limits at 50 turns. Has anyone else tried it?

Google is a great option. Have you seen their api library? Api monitoring will likely be a breeze.

Increase “live thinking” as assessment.

Anna Yeadell I think we need to look at this as a gradual shift over years. The writing in 1960s phd theses that I have read, and the work done is massively different to today, less literature referenced, more descriptive language even in science, more philosophy and less clipped. Using AI and then raising the bar through new more rigorous expectations in grade descriptors might be the only way forward? We need to keep the philosophical thought. Training in that is key.

I call bullshit on that one: AI is opening up my brain and buying me more time to spend outside and with family. Although i am a sample size of 1…

With the greatest respect, l think academicians, need to change the way they Assess.. For example a large part of the coursework should be practical where pupils have to answer questions on what they have written and that should form a large part of the grade. AI is here to stay so it's time to hear some creative solutions.

I have been using AI last 2 months 10 hours a day... I still can make coffee and boil eggs...We will see next month

I am sorry to say that I my experience, this was the case even before AI.

This could have been in the movie idiocracy 😂 of course it would be Elon that figures out to use human power. This is gold

Sharmin Shabnam Rahman, PhD, FHEA, CMBE, FCMA, CGMA.

15 milhões de méritos.

'Flash' unfortunate naming choice

Scary

Learn more on my page or my github

article link?

This is straight out of a Black Mirror episode.

This is hilarious, and I laughed out loud on the Altman character.

What I hate most is we are grading AI, not students. This fact makes our evaluation nonsense.

I hear the worry. But watch D. Scott Phoenix's TED talk — he argues we're not being replaced, we're going to merge. Evolutionary biology, not science fiction. It's actually less scary than the spin class version. https://www.ted.com/talks/d_scott_phoenix_why_humans_should_merge_with_ai

Very relevant. The easiest thing to copy today: UI + feature layer. The hardest thing to copy: Years of customer trust, usage patterns, internal workflows, and real-world distribution. A strong AI product isn’t just model + interface. It’s: AI + proprietary context + workflow integration + user trust + operational scale That’s where defensibility compounds. Great visual Rubén Domínguez Ibar