The Aboard Podcast
Is Jev the new shape of AI?
LLMs are often far too powerful for the small tasks they’re asked to complete. Is the more-contained Jev the alternative? On this week’s episode, Paul and Rich read a question from tech writer John Pavlus about Jev, a new model from Typesafe AI that’s been a major topic of discussion in programming circles recently. LLMs, Rich suggests, are often like using a power riding mower to cut the grass of your tiny lawn. As Paul breaks down Jev and others like it into layman’s terms, they debate whether this kind of technology is the future—or if it’ll just be one more thing absorbed by the AI giants.
Show notes
Links from this episode
- John Pavlus
- “Introducing System One Models & Jev” by Diogo Almeida
- The Jevons paradox
- That’s ask@aboard.com!
Transcript
Read along
The transcript for this episode is coming soon. Until then, watch it on YouTube with captions.
More episodes
All episodesCharity Majors: embrace your computer’s chaos
Discussing observability in the AI era with the co-founder and CTO of Honeycomb.
AI companies shouldn’t police themselves
Tech execs are sounding alarms about the dangerously fast pace of LLM development. Will any regulators stop them?
AI versus…talking to people?
Problems with your boss? Why not fire up Claude and hit HR with a lengthy legalistic complaint?