phillipsmg

A pretty reasonable statement to start an incredibly tough gig.

I don't see how he can turn this around, but kudos for trying and communicating openly about the short term plan.

How do you think the LLM -> API tool should work?

This is really neat. Barrier to entry is super low. Congrats!

I'd love the same but for meal plans, if you fancy building something else

Hard for someone else to answer a question about you. I still love Hacker News and check it most days

You may be thinking of https://learn-anything.xyz/

This just seems like redefining the claim. If the meshing hypothesis isn't true, doesn't that make the theory pointless?

The way this gets used in schools is almost always with respect to the meshing hypothesis

No union required for dumb decisions. But in this case, it does seem it may contribute

Definitely agree with the part about education and getting kids to learn!

Not sure about the comparison with beta-amyloid though. There are good reasons to think it is involved with Alzheimer's - genetics, histology, etc.

And I don't think many researchers thought that would be the problem solved. It was just a very logical target in tackling a very challenging disease (like all neurodegenerative diseases, basically none of which are cured).

Mainly following relevant accounts on Twitter.

This field is moving really fast right now, and so it's a little hard to keep track of all 'AI'. I'd recommend finding some researchers in the subfield of AI you're most interested in and following them on Twitter

website says pytorch i think

I would love for people to work on open science. I think it's near criminal that tax-payer funded science sits behind paywalls.

It's gradually changing, but the publishing companies make an extraordinary amount of money by acting as a bogus distribution system. They've managed to sit between academics and readers and charge both sides for doing nothing.

That being said, this isn't so much a tooling problem as a social one..

We're building a learning platform that adapts to your knowledge level and interests to tailor content to you. So far, we've built a knowledge graph of concepts and dependencies for ML, meaning you can find the fastest path to your goal, saving you time (https://app.learney.me/). We plan to grow out our knowledge graph and questions with a combination of expert input and community contribution.

We've just launched the first version of a major feature - questions tailored to your interests and background sent via Slack initially. Our long term goal is to build an intelligent tutoring system for tech subjects. You can sign up at the link!