OGRiki retweeted
To make your agent produce high-quality output, equip it with a way to check its work. Tests are part of that.
Here are some of the ones I think are worth investing in.
- Write end-to-end tests that simulate real user flows and act as the ground truth
- Use property-based testing to specify what should never happen and generate thousands of test cases that attempt to break those constraints
- If replacing a system, compare the old and new versions using randomly selected inputs
- Require tests to run quickly and deterministically, which reinforces the agent to retry when the loop is slow or unreliable (the agent may learn to retry, not fix the problem)
Example tests say what should happen; properties say what must not.
OGRiki retweeted
Practical Loop Engineering 👌
addyo.substack.com/p/practic…
@addyosmani covers goals, loops, and why not delegate your judgment
OGRiki retweeted
We've been working hard to get Networking up and running, so what better place to start than sharing that progress with the rest of our troops!
Join us in one hour (5pm GMT) for the Live Premiere! 🔴
youtube.com/watch?v=XeGGDN7I…
#hogsofwar #onlinemultiplayer #gamedev #premiere
OGRiki retweeted
SkillsBench: a benchmark that measures how well models use skills across a range of tasks. And it's so new that models aren't RL'ed for it yet!
skillsbench.ai/
ALT Graph of skillsbench
OGRiki retweeted
Nice reflective post from @slicknet :
"What we lose when we stop coding"
newsletter.humanwhocodes.com…
He cautions against over-reliance on AI coding agents for 100% of tasks and suggests spending a short daily session on unassisted coding each day.
OGRiki retweeted
🔥🚨BREAKING: A petri dish filled with human brain cells just learned to how to play DOOM
Scientists at Cortical Labs have successfully trained a cluster of approximately 200,000 living human neurons, grown on a microelectrode array chip, to play the classic 3D video game DOOM.
OGRiki retweeted
🦀 If you want to level up as a Rust engineer: stop reading only blog-level abstractions and start reading real code under stress.
Great example: profiling memory pressure, tracing lock contention, fixing the with data-oriented design.
mnt.io/articles/about-memory…
#rust #rustlang
OGRiki retweeted
Yesterday I gave the first talk in our 6-part series about building agents with agent-framework. Catch up with...
Recording:
youtube.com/watch?v=I4vCp9cp…
Slides:
aka.ms/pythonagents/slides/b…
Annotated write-up:
github.com/Azure-Samples/pyt…
ALT Slide describing an agent
OGRiki retweeted
Unreal Tournament 2004 is now available for free.
...well, I mean, you probably could've grabbed a GOG offline installer out the back of a truck earlier, but NOW it's Official/Legal and has the in-development community patch built in for ongoing support. oldunreal.com/downloads/ut20…
OGRiki retweeted
Did FPS games peak in the 90s? I guess the newer ones have better graphics, but do they really "feel" better in terms of gameplay and fun?
Unreal Tournament (Epic Games, 1999) was one of the best of its kind. A favorite for our LAN sessions back in the day. Carrying a 21" mintor over to your friend's house felt a little less heavy when you knew you'd be playing Unreal later that day...
Exellent and colorful graphics (even on PCs that weren't high end, like those you needed for Quake III Arena for example), great sound, and your typical capture the flag or deathmatches - I mean, what more did you need?
Plenty of "Game of the Year" nominations (with some wins), the highest critcal acclaim, and massive sales numbers. I think we can all agree that Unreal Tournament belongs in the Hall of Fame of FPS games. Or maybe I am just getting old and gloryfing things from the past too much.
OGRiki retweeted
Windows is the worst digital product in today’s world.
Bloated, inconsistent, unreliable.
It’s crazy how we gamers have no other choice because there isn’t any viable alternative.
OGRiki retweeted
Here is how I do and don't use agents, idk who this will help but its worth spelling out my preferences and why:
- I tell the agent to code how I would do it
- If the language is one I am _very_ familiar with I feel comfortable getting it to generate very good idiomatic code that is indistinguishable from my own and doing large refactors
- If the language is one I'm not comfortable with, I keep the pull request under 100-200 lines of code for the reviewers sanity since I can't discern the nuance of good/versus bad code
- ALWAYS read/self review the code before opening the PR, the onus is on the AI wielder to make sure the code is up to par with what they would do themselves before inflicting their teammates
- never auto open PRs because of ^
- if you do all the above you can avoid slop and not annoy your peers
my name is jessie frazelle and i have not touched code in an editor since october.
OGRiki retweeted
I recorded a 10-min video showing how I use agent skills in GitHub Copilot to turn weekly office hours into structured, timestamp'ed write-ups:
youtube.com/live/l_sdnLWNwUc
ALT Screenshot of GitHub Copilot using skills to generate write-up
