I gave 4.6, 4.7 and GPT 5.5 the same prompt and task to reverse engineer a collection of sample vector files from an obscure Amiga CAD program and create a detailed txt specification and a python converter that converts to SVG and produce a report so I can visually verify.
4.6 did very well. 90% perfect on first try, got to 100% with just a few followups.
4.7 failed horribly. First produced garbage output and claimed it was done, admitted it did that when called out, proceeded to work at it a lot longer and then IT GAVE UP.
GPT 5.5 codex was shockingly good. Achieved 90% perfect on first try in about a fourth of the time. Got to 100% faster and with fewer follow-ups.
Just tried with DeepSeek V4 Pro with OpenCode. It didn't do great. First attempt produced somewhat correct drawings for some of the original samples, but most were just a spaghetti messs of lines. Some prodding got it to do a little better, but still not right. A third prod and it went down a wild rabbit hole and was much worse. I gave up.
I also tried GLM 5.1, it's first attempt was such a disaster I didn't bother working with it any further. It also took by far the longest and wasted a bunch of time/tokens trying to find other converters online (and failing) instead of just reverse engineering the format from the sample files given.
Sure, with other changes. You would still have to:
* Select (or repair) “good” pairs that didn’t have poor echo, impedance, noise, etc. characteristics.
* Not compand the resulting PCM sample with mu/alaw companding.
* Provide an atomic bearer channel from the telephone switch to the ISP that was at least 112kbps.
* Develop a modem encoding scheme that could take advantage of the extra bandwidth.
At this point, you’ve reinvented a really crappy version of DSL that could be user-signaled to call different networks (i.e., phone numbers.)
The other option, the one that people selected, was to use the copper facilities but bypass the phone switch and its associated audio and in-band signaling shenanigans. That was/is xDSL.
Bumping ADC to 16bit will remove some of quantization noise, but still leave you with inherent noise of the transmission medium and power limit (FCC/AT&T).
No because you still have to modulate the signal to get it to/from a house, and modulation schemes are where you squeeze maximum bits per second into available analog bandwidth within the limits imposed by Claude Shannon.
The article didn't talk about modulation schemes and I wish it had.
I’ve made many a trip to the git repo to dig up older versions and copy down the old version ebuilds into my own overlay: https://github.com/gentoo/gentoo
Thankfully kitchen sodium is in compound form, and thus not likely to react violently with water. In this context, the properties of pure metallic sodium are relevant because it would need to be handled in manufacturing. Kitchen salt is more commonly mined or extracted, requiring minimal to no handling of pure metallic sodium.
4.6 did very well. 90% perfect on first try, got to 100% with just a few followups. 4.7 failed horribly. First produced garbage output and claimed it was done, admitted it did that when called out, proceeded to work at it a lot longer and then IT GAVE UP. GPT 5.5 codex was shockingly good. Achieved 90% perfect on first try in about a fourth of the time. Got to 100% faster and with fewer follow-ups.
I’m impressed.
reply