the first edition of the https://t.co/MeVtBI4lT6 course just finished.
it's difficult to find words to express how grateful I am to all the students who trusted us enough to join, and then enjoyed the course so much that they made it to the end.
now, seeing all the posts with kind words about the experience and the results they're getting is so rewarding and heartwarming.
none of this would have been possible without @theo@techgirl1908, @kentcdodds and @johnlindquist. I still can't believe how much trust you put in me, guys. I'm thankful for all the insights, experience and work you put into this project.
MEGA was undoubtedly the toughest project I've done so far, and there are many reasons for that. at the same time, it brings me the most satisfaction and motivates me for what comes next.
it also showed me that it's not only the way we work that is changing, but also how we learn and what's important now. I think in the coming months I'll be pushing the ideas that came to me over nearly two years of preparing this project.
meanwhile, I'm looking forward to seeing stories of how the MEGA experience translates into the daily work, products and initiatives of our students.
thanks again!
I earned the MEGA Developer certificate!
I enlisted a cohort of 5 other teammates, technical and non-technical, that our VP trusted in me would gain from the experience. (no pressure)
I had some love and love-less takes around the platform experience of @megadevhq, and at times, it was tough to keep motivating our cohort to keep up with the challenges.
But every week I put up PRs around the lessons learned, and scheduled our retrospective to reflect on the practical take aways we could bring to our work and others on the team.
We are already AI pilled, but if anyone is considering these lessons in the future, I highly suggest taking a similar approach if you want to bring more AI advocacy into your organization.
The challenges are easy to ignore, but they're designed to get out what you put in them. If you don't want to understand a challenge, your agent will do it for you. If you want to understand what your agent did, you'll be more of a MEGA Dev and less a MEAT Proxy 😂
I want to thank @theo for combining an in-depth understanding with his trademark delivery. Week 1 was by far the most impactful for everyone on the team.
Thank you @techgirl1908 for the introductions to Mastra and particularly classifications and evals; which I am now using in our pilot of a team performance coaching agent.
Thank you @kentcdodds for always seeing the forest through the trees; and the tea leaves so to speak when it comes to the role of engineers against industry changes. People are slowly catching up to how great Kody is!
Thank you @johnlindquist for bringing it home with ideas on how to further improve our own autonomous systems and orchestrations, but also thinking differently about harnesses––as I build my own.
And finally @_overment and the entire MEGA team, @joelhooks, @greg_rog, @BartekRycharski, @vojta_holik, @matsugfx, @sulich_olaf, Martyna Ruta, Anna Piskorz-Nowak, Kate Cwiklinska, Ann Golebiewska, Michal Jablonski, Mikołaj Banaszek, and anyone I missed.
Really appreciate everyone's contribution to this experience both here, in the discord, and behind the scenes 🙏
there’s something oddly satisfying about working on a fresh laptop without most of your usual apps and settings.
it’s easier to spot your bad habits and see what you actually need.
i made my macbook trackpad talk using only the vibrations from its haptic motor
my speakers are muted, every sound you hear is coming from the trackpad itself
open source, try it here: https://t.co/MPEKltkFxO
Andrej Karpathy: "A 1B model can be much more effective than most current LLMs.
90% of current models waste a ton of capacity and money on remembering stuff that doesn't matter."
true that.
and that comes down to knowing how to use ai to explore unknown unknowns and work well beyond your circle of competence.
it’s not just about choosing a tech stack, like ‘rust vs typescript’. it’s about getting rid of the labels that define our roles.
Agent platform of 2027.
This is the end game. This is the blueprint for the one and only application layer software that will exist. Every other software will effectively become databases for this layer, the everything agentic operating system for the million agent swarm to work on top of.
This is obviously the end game vision for the application layer of ChatGPT and Claude. I have reasons to believe that an independent company will own this product layer. But most ADE, "enterprise agent platform", or vertical saas companies, are focused on the wrong things far too narrowly, not technically competent, and do not have the product / technical vision to see this inevitable future.
The correct agent platform that wins will function like the base of a garden, from which customers will plant their unique "seeds" to grow their own plants and trees, managed and evolved by agents, humans, and signals from the sensors of the business. This presents a clearer picture than "factory of a factory" imo.