- Now that LLMs are so ubiquitous, I really enjoy hearing about how people are using them and the workflows they are creating - a lot of the kinks have been worked out, so now I am willing to reap the rewards and figure out how to adapt them into my life.
- Every time I hear a podcast like this I want to get back into sending off agents to work on my side projects while I do homework, and I will do it for a bit, but then I run out of time to review the agent’s pull requests, and I am stuck right back at the beginning.
- I feel that in order to really benefit from agents working in the background, you need to be able to spend some time every morning before work and every evening after to create a plan, set the agents to implement the plans, and then review them. Once all my random (around the house) side projects are complete, I am looking forward to taking part in the experience.
- Another thing I can never grasp is how people can afford the inference cost of running Claude Code all day - I get by through using the GitHub Copilot requests that are included in the Student Developer Pack, plus the free Gemini allotment for starting projects.
- The cost definitely makes me want to look more into self hosting an agentic workflow, since the power is pretty cheap here in Canada.
- Finally, I enjoyed hearing a little bit about the behind the scenes of Tailscale - I was also not expecting that. As someone who has just set up a home server cluster and NAS (which is accessible to the outside world through Tailscale) this was a fitting time.
podcast the-changelog programming tech open-source ai