| 15:58 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2025-06-13#1117117 << 'ocean of 74xxxx' aint imho very interesting other than from purely archaeological pov. |
| 15:58 |
dulapbot |
Logged on 2025-06-13 00:30:56 crtdaydreams: I'd be interested to know why you didn't pursue it asciilifeform |
| 15:59 |
asciilifeform |
(i.e. same reason as wai not very interested in mit cadr ) |
| 15:59 |
dulapbot |
Logged on 2020-08-06 19:20:48 asciilifeform: almost certainly shares the main defect of it tho ( i.e. much of the cadr os is intimately tied up with the kludges required by period silicon ) |
| 15:59 |
asciilifeform |
( see e.g. elsewhere ) |
| 15:59 |
dulapbot |
Logged on 2020-02-11 16:30:18 asciilifeform: 'cadr' was designed to be buildable with the off-the-shelf discrete semiconductors ('bitslice') of its day, and as a consequence extremely baroque. 3600 inherited this gnarl. |
| 15:59 |
dulapbot |
Logged on 2020-02-11 16:29:26 asciilifeform: that being said, the 'cadr' and '3600' series machines are likely the ~last~ thing a would-be designer of 'lisp cpu' oughta be looking at; |
| 16:00 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2025-06-13#1117118 << various things, thus far, but very little that imho anyone would want to read about |
| 16:00 |
dulapbot |
Logged on 2025-06-13 00:31:18 crtdaydreams: You've been off work for a while, too, what have you been up to? |
| 16:02 |
asciilifeform |
http://logs.nosuchlabs.com/log/asciilifeform/2025-06-13#1117113 << actually tried that one (& similar) for digging through chip datashits. verdict: worse than useless |
| 16:02 |
dulapbot |
Logged on 2025-06-13 00:26:41 crtdaydreams: I'd spent a little bit of time studying the schematics, and using an LLM assistant to try and understand it |
| 16:02 |
* |
asciilifeform so far asciilifeform's verdict on 100% of gptism |
| 16:04 |
asciilifeform |
'hey bot, what bits should i set in what register on rockchip 3128 to switch video output to lvds?' bot proceeds to give detailed but wholly worthless answer cooked from soup of datashits for other chips, incl. from other vendors, etc |
| 16:07 |
asciilifeform |
naggum's 'imponade'. |
| |
~ 5 hours 58 minutes ~ |
| 22:05 |
mats |
o3 is pretty good, but you have to hold its hand through tasks it probably doesn't have a lot of training data on |
| 22:10 |
mats |
o3 + deepresearch has been helpful for trawling dozens of search engine results about, for example, legitimate admin tools, lateral movement techniques, and the artifacts they create |
| 22:12 |
mats |
a few queries and ten mins while it searches and sifts saves hours |