Olisiko täällä henkilöä, joka haluaisi auttaa käymään läpi KiCad projektia, jonka tein täysin Claudella + Konnect piirilevysuunnittelu-tekoäly-työkalulla? Itse olen aika amatööri mitä tulee piirilevyjen suunnitteluun, enkä juuri ymmärrä siitä sen kummemmin, ja haluaisin vain varmistuksen, että onko tämä käyttökelpoinen millään muotoa? Ja mitä muutoksia tarvitsee jotta se olisi käyttökelpoinen?

Ja jos sinua triggeröi AI:n käyttö, niin voit jättää suosiolla kommentoimatta, tiedän kyllä että se ei ole kaikkien makuun.

Mutta tämän tarkoitus olisi toimittaa harrasteauton digimittariston virtalähteen virkaa, sekä muutamia lisävarusteita liittyen aihepiiriin

Ja työkalu jota käytin Claude Codessa:
https://github.com/mixelpixx/Konnect

No matter how robust ML models become, a statistical model cannot replace formal verification. Statistical models are non-deterministic by nature.

AI has been developing computer chips for years now.

Literally just doing a copy and paste of the reference design.

This is impressive but it’s not really representative of what is needed to design new circuits from scratch. There are so many designs for simple dev boards online that they are overrepresented in the training data.

This is basically typing out a hello world program.

Claude Fable 5 has demonstrated that AI can handle an entire PCB design and manufacturing process, although the results aren’t perfect.

A Reddit user created a working custom board for a Raspberry Pi Pico, display, and buttons without manually editing or verifying the design. The project involved several mistakes and cost around $150 for five assembled boards, while the total API usage reached roughly $450. Despite the cost, the finished board worked as intended.

Read more here :https://wccftech.com/claude-fable-5-ai-custom-pcb-design-demonstration/

Instead of putting its control computers into early retirement, CERN is migrating over 2,200 systems from IBM.

The European Organization for Nuclear Research (CERN), famous for housing the Large Hadron Collider, has unveiled a complex plan to transition its critical embedded systems to Debian Linux.

The decision stems from Red Hat Enterprise Linux’s (RHEL) shift toward newer CPU architectures, which utilizes compiler flags that drop support for older 64-bit processors. Notably, RHEL is owned by IBM. For CERN, complying with RHEL’s latest updates would mean instantly obsoleting 47% of its custom-built embedded control computers, with another 17% threatened by upcoming releases.

Rather than sending thousands of functional, highly specialized computers to the landfill, engineers chose to adopt Debian, an open-source operating system renowned for its broad, legacy hardware support.

While switching to Debian saves millions in hardware replacement costs, it presents its own set of high-stakes hurdles. CERN’s custom hardware requires highly specific kernel drivers and system images, and unlike RHEL, Debian lacks automated package-building toolsets natively. To manage the migration, engineers are evaluating specialized provisioning tools like Debusine to orchestrate the complex deployment. This major shift highlights a growing dilemma in modern computing: the delicate balance between pushing for bleeding-edge software optimization and preserving functional, expensive legacy infrastructure that keeps groundbreaking scientific research running.

source: Maruccia, A. (2026). CERN is moving more than 2,200 specialized computers from RHEL to Debian. TechSpot.

corollary: software eventually works, hardware eventually fails