Journal Week 08 — Keeping Up with AI Changes

Brief

Hi, no blockers… bye! Nah, just kidding. This week was somewhat interesting. Last week, OpenAI released its GPT-6 Sol and Luna models. Personally, I think they have become a game changer while working, especially since Luna costs a hundredth of what Astra does.

What’s new?

This week, I finally got a more concrete internship project. I am joining one that was already in progress. Broadly speaking, the project is about creating a Jira-triggered automation for handling tickets.

Current development?

I would like to say that the project is already super advanced, but nope. We are still in a research phase to determine which option is best and which has lower resource requirements, so we can adapt it and scale it if necessary.

Any other things?

I will have to give my TED Talk soon, and honestly, I am a little nervous about it. The topic is supposed to be open, but it needs to include a call to action, and I do not know what topic to choose for that yet.

Also, because of the freedom the office gives me, I have been able to spend more time with my friends. I do not have to stay at my desk all the time, so I visit them on other floors often. It makes coding and working on things much more enjoyable.

Learning and progress

This week, I have been strengthening my project-planning skills. Honestly, I think planning is the most important part of development. If I have learned anything from hackathons, it is that spending 60% of the time planning is often more effective than spending 60% of it developing.

Next steps

• Start planning my internship project and do more research.

• Start preparing my TED Talk.

Reflections

Having more models to choose from, along with Oracle’s constant reminders about my AI usage costs, has made me more aware of my overall impact on AI costs. It may not be a huge amount, but I sometimes think that a significant part of AI costs could be avoided with stronger, more deliberate manual planning.