That fits with my experience. I used Claude Code to put together a pretty complex CRUD app and it worked quite well. I prompted it to write the code for the analysis worker, and it produced some quite awful code with subtle race conditions which would periodically crash the worker and hang the job.
On the plus side, I got to see first-hand how Postgres handles deadlocks and read up on how to avoid them.
Or more cynically they reach their level of competence, go one level further and stay there to keep them from ruining the productivity of the people doing the work...
"I was wondering if you could go over the details of the resolution we discussed on DATE? I'm giving it consideration and just want to make sure I fully understand. On reflection I feel like it could be a great opportunity. Could I put it on my resume'?"
Sometimes they stitch themselves up and it's glorious when it happens.
Source: I was there. Person who handled this is a friend of mine.
They weren't milking it for attention. At EMF - at least early on - they were deliberately downplaying it to avoid causing a panic until the risk was known. Unfortunately when Atomicmaya's toot[1] dropped, they felt like they had to respond to squash any potential speculation or rumours. (the fear was someone would hear "orphan source" or "nuclear material" and think "Goiana incident, repeat of" -- Goiana was a much stronger caesium-137 gamma source).
Photos of the unit were being circulated privately in case there were more (the donator's identity was unknown at this point). EMF later announced in closing -- and you can see this on the recording [4] -- that they'd like to know if there were only two.
I've done basic risk assessment in a volunteer role and when there were unknowns, we took the path of assuming the worst, and planning for the best until we had more information. I can't really judge Tryst or anyone else for doing the same. In this case the worst-case scenario was a kid or teen at the camp buying it and taking it apart in their tent, and making the source material airborne.
Several people in the Furry Village google were trying to find information on the MIC based on the photos, and at the time we all found nothing.
The thing also very industrial which probably amped up the risk profile a bit further in peoples' minds, because it wasn't an obvious, recognisable, smoke detector.
This evening I googled the part number and sure enough, it's a very spicy (compared to modern ionisation detectors, about 10x the amount of Am241) early-generation smoke detector, and the Am241 is encased in gold. [2] [3]
TLDR: It's low risk, but that wasn't known at the time.
MGR was a later one - it could run (slowly) with an unmodified machine but really needed a VIDPAL to be used best. VIDPAL was a replacement PAL which enabled user-mode access to the framebuffer.
And probably the only real at-scale deployment of the Western Electric 32000 (Bellmac 32) processor. Sadly that chip wasn't very popular.
As LeoPanthera said above, Bell's exchange for the ability to make PCs was a bit of a bad swap in retrospect. It's pretty easy to see they were going for some kind of convergence of computing and telecoms (the 3B1 has an internal 1200 Baud modem) but they didn't do well on that bet.
It's really a shame how little of that software has survived. I think I have a complete Foundation Set disk set, some others I wrote from ImageDisk images, and that's about it.
UNIX software in general seems pretty thin on the ground. The OSes are out there (even rarities like Interactive Unix 4.1 - though not the later patches e.g. FDISK 2GB) but software? Hen's teeth.
ACCELL Unify would be interesting to find, it's the software General Instrument used to write the front-end for their cable TV headends.
Ah, that's neat - it'd be nice to have it for SCO or Interactive Unix on x86. Especially Interactive, as that's the one I have to try and fix (Y2K20 issues).
The VM support was the killer feature. The 3B1 has a hardware (discrete logic!) MMU, a custom designed thing completely unlike the Motorola 68451. It's an odd beast and took a lot of work (by several people, not just myself; check the git commit logs and CREDITS) to get it right.
reply