Who we talked to, what we learned🤓
New podcast with Mike Bird, has dropped (Youtube, Spotify, Apple, and Substack).
Once again, the Jetsons was referenced, Mike takes us through some interesting use cases and how opensource is catching up.
Also available on Youtube
and I had a chat on the future of open source and community building on his podcast (see here)
In it we spoke about hype cycles, pragmatism, a Predibase survey and a bit more. Check out his channel and Substack for highly in depth coverage of legal and AI crossovers.
As someone who works (largely) alone, at home. I find having that the lack of designated “work buddy” (read- therapist, confidante,”ride or die” or occasional conscious) makes falling down mental rabbit holes a lot easier.
Emily Witko (who recorded an episode with me recently) said it best on their Linkedin with “Throughout the day I can be seen sharing memes and asking someone to think things through with me more than once or twice 😬 but it's not fair to them if I'm just a distraction 😅
I’ve been using the Rubber Duck assistant throughout my day and have found it excellent.
I’ve spoken before about my absolute love of goblin.tools this is a similar vibe.
🤘 (un)supervised learning is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
What are you building, locally?
I took a little adventure into the LocalLLaMA subreddit the other day, curious about what all these clever folks are up to with their use of local LLMs.
Things like home automation, data sorting and categorising, Jarvis (because everyone is building Jarvis ;))
Hanging out in the LocalLLaMA subreddit is like peeking into the future. And it’s helped me recognise a few things about (un)Supervised Learning.
It’s made me recognise that we need a home.
It’s made me recognise what (un)supervised learning isn’t
It’s not a newsletter with “the top AI tools you have to try”
It’s not somewhere you have to be an expert.
It’s not just to talk about the newest thing.
I’ve created a form where folks interested can nominate their level of interest and participation. (made with an opensource Airtable 🤗)
Here’s the form again
Feel free to share it around.
The “Meat” (of the post..)
I am tired of every day, feeling like an idiot because I don’t have a machine learning degree.
everyone is a genius. but if you judge a fish by its ability to climb a tree, it will live its whole life believing that it is stupid.- Einstein- maybe
I get DM every day from folks curious about the best resources for X or Y, who they can reach out to for more info etc.
I’m not a repository of knowledge. But I believe the community can be!
Soon, I’m wanting to try out a platform called Superwave
Below, you’ll see a message I sent to a friend, Claudia of NoCodeOps, who suggested Superwave in the first place.
As you can see, a lot of questions.
So what do I envision for (un)supervised Learning?
Collaboration- not “competition”
@unsupervisedlearningpodMy (so far) short exposure to OS culture has cemented a mindset shift- that when working towards a similar, lofty goal, whether in development or broader transformations, there is no competition. The ego-driven idealogy that we have to compete, outperform and "crush" competition is something I see myself rejecting as I go deeper in the opensource rabbit hole. Tying into the above ^ by default, it is much easier than I expected to reach the "right" people. Aka the people with a wealth of subject matter knowledge, a large following, accolades etc. #opensource #foss #github #llm #ai #wfh
Tiktok failed to load.
Enable 3rd party cookies or use another browser
Peer learning
Hackathons for “non-tech” people
& the list goes on and on
Questions I put to you
Hot topic recently- is natural language (text based) the best way to prompt? - What other ways have you found effective?
Why local?
On a more concerning note, I saw this alarming article in Wired about “AI girlfriends” and the massive data privacy issues they raise.
It's a reminder that not all AI is created with our best interests in mind. (tbh lean towards thinking most *is*)
These kinds of privacy concerns make me wonder if developing and utilising local AI solutions will become more important than ever.
When our data stays on our own machines, or within trusted local networks, we have a greater degree of control over who has access to it.
Of course, this isn't a bulletproof solution, but it could be a significant step towards safer and more responsible AI use.
"Do it Safely" Over "Don't Do it"
I think we need to shift the conversation towards educating ourselves and others about safe AI practices.
That means understanding how these models work, being critical of data collection methods, and advocating for greater transparency and accountability within the industry.
A video I really really enjoyed recently was shared by my friend at Jeezai on Andrej Karparthy and I believe this kind of knowledge sharing is how to REALLY scale innovation. Less “trust me bro”, more “here’s why”.
That was a lot 🥸
Ways to help out
Share 🤘 (un)supervised learning. This post, this form
Recommend someone for the pod- Have someone in mind who’s doing cool things across AI or opensource? Send them this way!
Follow and rate and all that jazz- Spotify, Youtube and Apple podcasts are a learning curve but I know the algo beast loves a rating, yours helps!
Stay curious,
Renee
🤘 (un)supervised learning is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
