e12e
Joined 14 years ago
Karma
15,425
Submissions
9,976
Member since
Jun 13
Recent submissions
Comment · 18 hours ago
I call it uncanny valley effect for text [Ed: ops use of prose, rather than text is better formulated]. It's a little like with AI images - they started out with six fingers, but now they look "ok" - but slightly off.
Reading some AI text feels like looking at a Michael Murphy sculpture - makes perfect sense, until you start to look closely at the text, and gaps appear - and you start to realize there's no there, there.
Comment · 19 hours ago
I wonder if they will publish slides and talks from?:
Comment · 2 days ago
Agentic coding work or just agentic work?
Comment · 4 days ago
Comment · 4 days ago
On the other hand, ai agent harnesses work a bit like image based system - think: claude --resume
Comment · 4 days ago
Odd, the dictionary definition (here: webster's) is literally:
> guy: (...) person —used in plural to refer to the members of a group regardless of sex
Comment · 5 days ago
Thank you so much for sharing your insight - it's not much serious discussion out there on sane reasons for self hosting (that I've seen).
You say the big rig is 25k - and then the two sparks are about 1k each?
Electricity is for "just" the big rig?
Comment · 7 days ago
Not sure how something like:
https://dev.to/aws-heroes/scalable-sequence-for-postgresql-3...
Would interact with sharding?
Comment · 7 days ago
My second thought was it sounds like the colorless green ideas have slept furiously, so to speak.
My first thought was: omg, they've harvested the souls of a dozen insipid, self-centered genz' and trapped them in a machine, like The Finn living in an Aleph...
Comment · 8 days ago
> You can download it for macOS and Linux
Comment · 8 days ago
Wonder how this compares to clasp (common lisp c++ impl)?
Comment · 8 days ago
I know you're joking - but I think it will remain a challenge.
Quality and Security are emergent properties of software - especially when considering software-as-deployed in the real world.
And the problems with "good taste" and good architecture and data modelling - will probably leave ai code open to a lot of issues (Claude code cli issues is a great example of how bad this can get).
Having good architects in charge can help - but it requires a lot of discipline to keep the feature set small so that maintaining quality and security become feasible.
Comment · 9 days ago
I think AI might make reimplementatiion from first principles more viable (think: an Oberon os with Linux driver support; or: a full environment like STEPS, but for real use) - which might make it feasible to simplify the problem domain (firmware, driver, os, UX, compilers) - to the point that it becomes easier to create bug free domains/subsystems.
Comment · 9 days ago
Congratulations on the launch - looks interesting, and matches some thoughts I've had regarding herding LLMs in a sensible and secure manner.
What's the product? Can I buy anything right now?
A detailed I noted on permissions - looks like they're all prefixed with "always" - I think allow/deny/ask make more sense ?
Feel free to contact me at: [email protected]
meet.hn/city/no-Tromsø