pages tagged with #llms
on the future of software
the times they are a-changin'
It's been approximately six months since I've spent an entire day just editing software in IDE. Six months without highlighting text on a screen and asking an LLM, or rather six months before attaching a snippet of code as context for an LLM in a chat panel to modify. Six months from having a terminal window open and iterating back and forth as an Anthropic or Composer or Codecs iterates quickly on the code in front of me. Six months since I've done anything other than look at a Markdown file detailing exactly the changes that are going to happen to me as I create them. The future of software has been changing so rapidly that I have yet to fully grasp or think about what will happen.
[... 962 more words to read 📓 ]A super interesting blog post about what the future of fully agent-driven software might look like. It's essentially a breakdown of what an AI-pilled high throughput coding system might look like. Where the role of being an engineer is less about the code being written in the system and more about scaffolding the system to make it extremely amenable to agents to use. Like the scaffold to create software that you want. Some of the things that were extremely interesting and stuff I had not seen:
[... 454 more words to read 📓 ]something i’ve been struggling with personally is how much to offload to llms. on one hand — the technology is fascinating, powerful, and enables me to learn and do more than before. on the other hand — it dulls my technical skills and robs me of learning opportunities.
[... 448 more words to read 📓 ]this is an interesting piece by Dex Horthy of HumanLoop (an ai startup working in the LLMOps space). it mimics herokus “12 factors apps” (of which i was honestly unaware of, my inexperience is showing 🙈). i've listed some takeaways below!
[... 785 more words to read 📓 ]an interesting blogpost by Geoffrey Huntley showing how llms can be used to "decompile" applications with bundled / minified code (and in a following post) they show it can also be done for assembly. i think this is especially interesting (and concerning for the creators) of electron applications, as there's relatively accessible source code in languages and frameworks that llms are actively being trained on, though i imagine there will be a day that most compiled apps can be undone.
[... 358 more words to read 📓 ]a ux for creation
how can we design better interfaces for llms?
chat interfaces fundamentally limit how i interact with LLMs. they’re incredibly effective at showing the power of LLMs to a wide audience — specially in the days where agi felt more like fiction than the goal of the tech industry. and for that reason, they’ve become the de facto standard. but, in my opinion, flaws in their design hold me back from realizing the true creative potential of these models.
[... 874 more words to read 📓 ]