I'm releasing a tool to free your twitter bookmarks and make them searchable. It's still a WIP but the basic functionality works. It does involve a couple steps, some of which are slightly technical. Steps below
https://t.co/JQqoaYlVoo
@ben_m_somers@WayneVaughan hm interesting, thanks. they act dumb as in they fail to follow instructions and guidelines?
I'm thinking of using the models for a pilot I'm running in a prof's course in a month or so. If I find anything interesting with these models I'll report back
@MilksandMatcha@_gongy Coast starlight right? Sometimes National park service guides will be in the observation deck and give a spoken guide of the surroundings
@IaimforGOAT@arestlessrest@viemccoy always a joy to see you periodically recommend this
Also surprising that it still works despite the project effectively being in "set it and forget it" mode for a while now
the outcome of all this is still uncertain but there's mutual interest to continue my involvement there and branch out beyond giving workshops. So I'm excited but anxious to see where this will lead.
There's somewhat of an uphill battle to fight especially since I'm an outsider to higher ed and don't know how to navigate the structures there, but there's a lot of genuine enthusiasm among the faculty to adopt these tools and techniques around AI, so the right ingredients are here to do something special
little update on the current side quest I'm on. I've been exploring an untapped opportunity that I think I'm uniquely positioned to exploit around deploying what's possible with coding agents to fields outside the tech bubble.
I gave a talk recently to faculty and grad students at UCSB's Department of Economics (my alma mater) on an intro to coding agents.
@JeremyNguyenPhD Would be very interested to join the reading group! Iโve been reconnecting lately with some of my previous profs in the economics department and am likely going to be holding workshops for faculty and students in the near future
I'm releasing a tool to free your twitter bookmarks and make them searchable. It's still a WIP but the basic functionality works. It does involve a couple steps, some of which are slightly technical. Steps below
https://t.co/JQqoaYlVoo
@CloudflareDev I also provide a markdown prompt I've been using to give a more guided and structured experience for surfacing interesting connections in your bookmarks. It also runs a temporal analyses into how your interests and behavior patterns might've changed over time
@p_millerd Rss. Assembly lets you use a url of an audio file as the source. So you can actually use the raw audio urls contained in the rss feed. So first step is feed ChatGPT your rss feed and have it extract all the audio urls
@p_millerd But a good workflow is https://t.co/CoWXvzMe8Q and then feed the transcript to an llm to fix transcription errors, clean up filler words etc