I finally pulled the plug on @Learn_Awesome. A few reasons:
- Human curation will become super valuable in the age of fast-changing knowledge and AI-generated content and therefore, will suffer even more from the free-rider problem (curators do all the work without getting much help from the leechers).
- Want to rethink tagging, taxonomy and similarity search using vector embeddings.
- It now seems possible to automatically build interactivity with personalization into existing curated static content.
- I would now prefer building for small but engaged user groups rather than anonymous web visitors.
@tedcurran @sama@gdb@ilyasut@paulg@OpenAI This was before GPT-3 pricing was rolled out. Today, one can try the GPT-Neo hosted API by HuggingFace here: https://t.co/uKlhXjvwwo
2.7B param v/s 175B params for GPT-3. Looking up now, Google built a 1.6T param and China a 1.75T param since then. π²
@alec_helbling@sama@gdb@ilyasut@paulg@OpenAI Answer to second seems a definite yes, at least on topics where questions require problem-solving.
The first one requires building a deep knowledge graph and then detecting the kind of mistake the student is making.
Good news! ππ We can now receive sponsorships and contributions thanks to @github sponsors and @opencollect:
https://t.co/norY3btZbb
https://t.co/kxtRkUtYbu
Made a few improvements:
- Added a section "Do Good" so that it feels less narcissistic/self-indulging
- Life-stage (age) slider is now persisted in localStorage
Try it here & suggest improvements. Note: your data is NEVER sent to the server: https://t.co/rfwaCYvQL8
the quicker people build lifelong education into a parallel system you can seriously opt-in to in lieu of the mainline k-12-college-grad school legible education, the better. i am starting to feel like it is pretty mission critical; like education is getting profoundly bad
@ringo_ring Hi Alexandra, thanks for all your work.
Question: is there a way one can know the IPFS hash of papers? We curate r best research papers on various topics but using URLs or DOI strings or ISBNs will always involve middle-men and therefore get in the way of durable access.
People seem to treat their current amount of knowledge as a constant β sometimes almost a matter of identity. E.g. they'll say "I am a non-technical person" instead of "I haven't learned to program yet."
The rotten culture in the organizations is often propagated from the top to bottom.
Please counsel your staff, friends, whosoever to not shell out their hard-earned money on these apps, plenty of free, good resources are available. Check - @cs_pathshala, @Learn_Awesome.
Simply click on the Gitpod badge in the README to get started: https://t.co/tgfCFej3Dy
Here is our gitpod config if you are curious: https://t.co/f7eHyzdUOU
LearnAwesome's open-source code repo is now @gitpod ready! Simply visit a URL to get a full-fledged development environment with pre-loaded database, terminals for all needed processes, and a proper web URL to see your changes in action! π²π₯³π
See the README.
π₯³πNew features for course-creators!
You can now make your course invite-only. Create invite codes & define their limits. Most common reason is to require a payment for your course! πΈ
A course can now be run by multiple teachers with access to dashboard and chat rooms! π