Engineering and R&D | Cofounder & CTO at @recalllabs_ | AI, Machine Learning, and Distributed Systems | Collector of maps | 🇨🇦/acc
- Tweets 6,979
- Following 1,108
- Followers 39,550
- Likes 9,260
carsonfarmer retweeted
The more time I spend working with coding agents, the more convinced I am that they make software engineering even harder
We can do amazing things with them, but unlocking their full potential requires extraordinary discipline and knowledge
Bury your head in the ground at your own risk. I aim to not jump on any hype trains, but since Opus 4.6 and GPT-5.2 + the harnesses it was clear that these things can write code nearly as good as I can in my best language; better in other languages.
But we won't see non-devs push production code for a long time; possibly forever, if you ask me. Software still needs to be built in robust ways, and it's our profession to do this well, and use the new tools we have, which create new and interesting (+tough!) challenges
349
381
4,439
597,096
carsonfarmer retweeted
This is the way. Accept code on merit and ensure that a person takes accountability for it. Doesn’t matter if it was typed, chiseled, generated, or bit-flipped via magnetized needle on a chip.
If it’s good, it’s good. If it’s slop, it’s slop.
KDE is working on an official AI / LLM policy, and it reads like the rules of Fight Club.
In short, KDE’s AI policy:
1) Encourages using AI, as long as a human is kept “in the loop”.
2) But you can’t tell anyone that you used AI.
“Don’t disclose LLM usage”.
“Don’t add ‘Assisted-by: [some LLM]” (as is done in the Linux kernel).
“Nobody in KDE should know if you use an LLM”.
In other words: “Welcome to developing KDE with AI. The first rule of developing KDE with AI is: you do not talk about developing KDE with AI.”
invent.kde.org/plasma/plasma…
92
228
4,048
232,482
carsonfarmer retweeted
Thanks to everyone who joined the EF Protocol AMA with questions! Lots to dig into on zkEVM, post-quantum (PQ) Ethereum, and formal verification (FV).
Here's a guide to 24 relevant questions with answers, with direct links to each discussion.
1/12
Feel free to ask any questions about zkEVM, PQ, and FV. Our team can help provide more clarity on those topics!
2
13
28
3,534
carsonfarmer retweeted
Feel free to ask any questions about zkEVM, PQ, and FV. Our team can help provide more clarity on those topics!
The EF Protocol Cluster's 15th AMA begins in 24 hours
Ready to learn about Ethereum's potential futures?
Ask away!
→ Submit questions now: teddit.net/r/ethereum/commen…
→ Answers roll in on Sep 16 at 2pm UTC
2
11
3,401
carsonfarmer retweeted
We are seeing a new trend in submissions to @arxiv (and presumably to conferences and journals): Authors submitting papers whose contents they likely do not understand. 1/
190
356
2,779
662,051
Can I say something unpopular? I'm not building a 3D rendering of anything... so Astra is just kind of, good? But I'm not seeing the AGI... and I'm not convinced being good at blender is the right benchmark here y'all?
1
3
825
carsonfarmer retweeted
writing the code may be solved but making a good product sure as hell is not lol
3
5
33
2,930
Does anyone know of a cheap, fast, easy way to remove low-quality followers on x? I've got so many bots hanging around, I'm certain it is hurting my algorithm position. Codex is "manually" doing it, but hitting rate limits obviously.
2
3
720
carsonfarmer retweeted
The turbine rotation is even clearer if you cross-fade between each frame
Turns out you can extract the rotation of a wind turbine from a single pass of a satellite 🛰️
Each time a Copernicus Sentinel-2 satellite passes overhead it actually captures multiple images across different spectral bands.
These images are taken a second or so after each other, which means that fast things can be seen moving between each capture.
That's plenty of time to see a wind turbine rotating.
1
3
22
2,092
For anyone reading this and discovering Canadian tech, see also madeincanada.dev
Tech developers across Vancouver Island and Canada are creating Canadian alternatives for American tech giants as the 'buy Canadian' movement resurges.
cheknews.ca/vancouver-island…
723
Hot take: being good at prompt engineering is surprisingly close to being good at mechanism design.
You're basically designing the "rules of the game" so that you and your agents reliably converge on the outcome you want.
2
8
809
Obviously the prompt itself is only one part of this.
So once you have multiple agents, tools, memory, delegation, etc... you're basically designing the rules, constraints, information flows, and feedback loops that make good behavior emerge naturally.
1
101
An example from @aiworthusing podcast (podcasts.apple.com/ca/podcas…) about test rules in AGENTS.md is a good example of this.
It starts to feel a lot more like system design than "writing good instructions".
1
69
