As I was building out this Big Five assessment recently I realized it would be helpful to create Typescript and Python libraries allowing developers to hit the ML model directly. The app could provision them an API key and they would be on their way. I could picture in my mind the entirety of the minimal system architecture.
For a moment I felt The Friction. The friction bubbled up unconsciously from memories of implementing features in the before times. Most of all from the memories of debugging features in the before times. Of the time it took. Of the effort it took even just to prioritize them.
And just as quickly as The Friction showed itself, it washed away and was replaced with The Joy.
The joy that all I need to do is reach in with my specific knowledge and pull this feature out of the weights and harnesses and into existence and it will just work.
As I was updating the homepage of this website recently I kept coming back to my title of "Software Engineer". I realized this is a job I don't have anymore.
As the field of coding has changed over the past four years, I've grounded myself in the idea that jobs do not exist. Groups of tasks do. I suspect it will be important for many of us to recognize as much sooner or later. Along with humility and a sense of non-attachment.
I've learned that I may wake up on any given day and find another set of tasks I no longer need to do. Tools, systems, LLMs, agents, and harnesses continue to eat all my tasks.
Agentic Engineering resonates with me. But so too does something like Maestral Engineering which captures the power of creativity and curiosity when working with these tools.
For the time being I've updated my homepage to reflect "Agentic Engineering".
What stands out to me the most is how this will impact staff resourcing:
Before asking for more Headcount and resources, teams must demonstrate why they cannot get
what they want done using AI. What would this area look like if autonomous AI agents were
already part of the team? This question can lead to really fun discussions and projects.
I recently created this meta-prompt after applying what I learned from this
arXiv paper.
Exploring Expert Behavior Driven Prompting
The paper caught my attention as it seemed it could be a great resource for prompt design.
Just before reviewing it, I'd been struggling building a prompt with a strict set of output
requirements. One of them was to keep the output under 2000 words.
It kept creating well formatted outputs about 2500 words long.
This wasn't surprising as LLMs often "think" in terms of token length as supposed to word
length.
After summarizing the paper, I knew it could help. And when I was done it did. I was able to
get high-quality output in the 1400 word range.
Using the HTML version of the paper I asked Claude 3.7 Sonnet to summarize the paper for me.
I then followed up by exploring how the research could be applied to prompt design.
How would the learnings from this paper be used when designing system prompts for llm chat
assistants using the direct foundation model API (ex: openai, anthropic)
I then asked what a prompt would look like that incorporated the application of this
research.
Here's what I got.
Now I could use this prompt to make it even better.
The v2 prompt was created using ChatGPT 4.5.
GPT-4.5 created an improved result which I was able to use for another round of improvement.
And I ended up the the final prompt to use.
I used both Claude 3.7 Sonnet and ChatGPT 4.5 to evaluate all three versions of the prompt.
It was satisfying to see the v2 prompt score higher than the v1 prompt and the v3 prompt score
higher than the v2 prompt.
This entire process was a lot of fun. It's another example of how when we identify something
potentially useful, we can leverage LLMs to quickly validate its potential usefulness and
create new tools we can use.
I've torn this site down and built it back up a few times over the past ~2 decades.
I've never been quite happy with the tooling I've used. And so I've coded this site from
scratch using Cursor + Sonnet so it's come together fast and it's (more or less) just to my
liking.
I intend to use this Link Blog section to regularly share links in the style of
Simon Willison's Link Blog.