Monday, June 29
Daily News Stuff 29 June 2026
Suffering Sassafras Edition
Suffering Sassafras Edition
Top Story
- Tokenmaxxing is dead. Long live tokenmaxxing. (12 Grams of Carbon)
We have reached a tipping point where spending infinite amounts of money can sometimes make the result better rather than worse.
- AI coding costs are on track to match human payrolls. (Infoworld)
But you are still spending infinite amounts of money.
Tech News
- Why Wall Street thinks Micron is the next Nvidia. (Tech Crunch)
Wow, they really hate Nvidia over there.
- Chinese CPU maker Loongson has announced its upcoming 3C3000 CPU, with 16 cores and the performance of a... Don't know. (Tom's Hardware)
The company positions its higher-end 3C6000 against the low-end Intel Xeon 4314, or about half the speed of an AMD 16 core laptop CPU. Presumably it will be faster than last year's 3A6000, which was dogshit.
- The best time to buy a Mac Studio was a year ago. (WCCFTech)
The second-best time is in 2028 when the M7 Ultra CPU is expected to make an appearance. There may yet be an M5 Ultra, but that will drop into an otherwise unchanged MacStudio.
- If you want a slightly fancier watch that won't break the bank, Casio has two more options for you. (Notebook Check)
Still digital, and adding a pedometer and Bluetooth connectivity, starting at $80.
Musical Interlude
Disclaimer: Woof.
Posted by: Pixy Misa at
06:46 PM
| Comments (3)
| Add Comment
| Trackbacks (Suck)
Post contains 223 words, total size 3 kb.
1
I looked at your 3A6000 comparison to the Core 2 Quad Q6600, noting that they got 1400 and 1800, respectively, on the CPUMark.
Then I added the i7-14700K...and it gets nearly 52,000.
Then I added the i7-14700K...and it gets nearly 52,000.
Posted by: Rick C at Monday, June 29 2026 11:15 PM (GNgpN)
2
Hackaday occasionally links to projects who do mainboard replacements on those old Casio watches, giving them new features like BT. Interesting to see Casio playing catchup.
Example: https://hackaday.com/2026/06/23/a-custom-pcb-for-the-casio-g-shock/
"But the replacement board combined with the open source firmware brings new capabilities that far exceed anything the G-Shock was capable of originally. The upgraded watch now offers several applications, such as a pedometer and a number of games including simplified versions of Blackjack and Wordle. The watch can tell you the phase of the Moon, calculate sunrise and sunset, and display values pulled from the internal thermometer."
Example: https://hackaday.com/2026/06/23/a-custom-pcb-for-the-casio-g-shock/
"But the replacement board combined with the open source firmware brings new capabilities that far exceed anything the G-Shock was capable of originally. The upgraded watch now offers several applications, such as a pedometer and a number of games including simplified versions of Blackjack and Wordle. The watch can tell you the phase of the Moon, calculate sunrise and sunset, and display values pulled from the internal thermometer."
Posted by: Rick C at Monday, June 29 2026 11:17 PM (GNgpN)
3
I have been seeing stuff on twitter and linked here about the AI regulatory expansion, and have addled thoughts of my own.
This is not so much a Trump policy, as it is that these export control bureaucracies are their own thing, and basically prone to responding to things that they are told. The way not to get yourselves regulated under ITAR is to either obviously not be any sort of useful weapons system or terrorism viable weapons system, or to have made Yudkowsky, Amodei, and I think the Effective Altruists shut up years ago. Treating this as a necessarily Trump level policy change is to misunderstand how these bureaucracies think. (Misunderstanding how these bureaucracies think is an easy mistake to make, as most people have businesses that are not ITAR controlled, and even people who do might not fully understand the underlying thought. I don't fully understand the thought.)
Trump may be involved, but we don't know that, and he may actually be inclined to let these bureaucrats stand after he is involved in some discussions. Trump does not have one method. He can maybe be thought of as having half a dozen methods that would result in different answers to a given choice, and he picks an 'appropriate' one for each choice. This is maybe more ideological and more structured than his actual method, but it is maybe easier to explain to an 'intellectual' with one method.
There was a twit who basically called upon Americans to resist by action if this policy choice is what goes forward. I basically disagree that AI is an unusually bad result of ITAR policy. I also do not beleive that I am obligated to resist ITAR policy. Lots of second ammendment fans (who are not merely gun collectors) basically are willing to give up points on the first and second ammendment when it comes to ITAR. Long arms and pistols are important sticking points, but they aren't interested in fighting, say, destructive device rules. Internet publication of technical details is an export, and basically a lot of Americans are fine with not proliferating details of WMDs such as NBCs.
Anyway, I find it hilarious that there are people complaining that rolling stuff into these regulatory regimes is opaque, and is a matter of government picking winners and losers without clear deadlines. "My dudes, have you ever gotten a security clearance?" Non-diverse Americans with technical degrees often got into the Defense contracting/export controlled world because IT and Big Tech H1B slavery had driven them out. IT and Big Tech are the industries pushing on AI, which want access to the good services. Get rid of the H1B slaves, and get rid of the foreign intelligence assets, and then you can start working towards demonstrating compliance and spinning up the infrastructure to handle the verification. From this perspective, this regulatory scheme is entirely in keeping with some of Trump's expressed policy preferences.
The process is actually pretty well known. You read the government regulations. You find a contractor to give you training on what they mean. You try to comply. If the government gives you information, and then decides you mishandled it, they prosecute you, and send you to jail. You try to prove upfront to the government that you are in compliance. Then, if the government believes that you were dishonest, there are criminal prosecutions. Of course, there is a presumption of innocence in criminal prosecutions, so in theory they have to do a good job.
This is not so much a Trump policy, as it is that these export control bureaucracies are their own thing, and basically prone to responding to things that they are told. The way not to get yourselves regulated under ITAR is to either obviously not be any sort of useful weapons system or terrorism viable weapons system, or to have made Yudkowsky, Amodei, and I think the Effective Altruists shut up years ago. Treating this as a necessarily Trump level policy change is to misunderstand how these bureaucracies think. (Misunderstanding how these bureaucracies think is an easy mistake to make, as most people have businesses that are not ITAR controlled, and even people who do might not fully understand the underlying thought. I don't fully understand the thought.)
Trump may be involved, but we don't know that, and he may actually be inclined to let these bureaucrats stand after he is involved in some discussions. Trump does not have one method. He can maybe be thought of as having half a dozen methods that would result in different answers to a given choice, and he picks an 'appropriate' one for each choice. This is maybe more ideological and more structured than his actual method, but it is maybe easier to explain to an 'intellectual' with one method.
There was a twit who basically called upon Americans to resist by action if this policy choice is what goes forward. I basically disagree that AI is an unusually bad result of ITAR policy. I also do not beleive that I am obligated to resist ITAR policy. Lots of second ammendment fans (who are not merely gun collectors) basically are willing to give up points on the first and second ammendment when it comes to ITAR. Long arms and pistols are important sticking points, but they aren't interested in fighting, say, destructive device rules. Internet publication of technical details is an export, and basically a lot of Americans are fine with not proliferating details of WMDs such as NBCs.
Anyway, I find it hilarious that there are people complaining that rolling stuff into these regulatory regimes is opaque, and is a matter of government picking winners and losers without clear deadlines. "My dudes, have you ever gotten a security clearance?" Non-diverse Americans with technical degrees often got into the Defense contracting/export controlled world because IT and Big Tech H1B slavery had driven them out. IT and Big Tech are the industries pushing on AI, which want access to the good services. Get rid of the H1B slaves, and get rid of the foreign intelligence assets, and then you can start working towards demonstrating compliance and spinning up the infrastructure to handle the verification. From this perspective, this regulatory scheme is entirely in keeping with some of Trump's expressed policy preferences.
The process is actually pretty well known. You read the government regulations. You find a contractor to give you training on what they mean. You try to comply. If the government gives you information, and then decides you mishandled it, they prosecute you, and send you to jail. You try to prove upfront to the government that you are in compliance. Then, if the government believes that you were dishonest, there are criminal prosecutions. Of course, there is a presumption of innocence in criminal prosecutions, so in theory they have to do a good job.
Posted by: PatBuckman at Tuesday, June 30 2026 12:36 AM (s6adZ)
56kb generated in CPU 0.1092, elapsed 0.1851 seconds.
58 queries taking 0.1453 seconds, 368 records returned.
Powered by Minx 1.1.6c-pink.
58 queries taking 0.1453 seconds, 368 records returned.
Powered by Minx 1.1.6c-pink.









