Daily Tech News 13 September 2025

• Intel has announced two new desktop CPUs. Do not buy them. (Tom's Hardware)
  
  The first is the i5 120, a six core part that has the exact same specs as the Core i5 12400 from 2022, mostly because that's what it is. Admittedly not an awful part, particularly if you didn't run a workload that would make good use of the "efficiency" cores, because it didn't have any of those.
  
  The second is the i5 110, a six core part that has the exact same specs as the Core i5 10400 from 2020, mostly because... Yeah. The 10th generation chips didn't even have "efficiency" cores yet, so you're safe there. But you will need to find a five year old motherboard and DDR4 RAM for it, because none of this modern stuff will work.
  
  Oh, and it's 14nm.
  
  
  
• Arm has introduced four new CPUs. (Notebook Check)
  
  The C1 Ultra replaces the X925 (which replaced the X4) as Arm's new flagship mobile core.
  
  The C1 Premium replaces the A725 (latest in the same line as the A78, for example) as a sub-flagship core.
  
  The C1 Pro also replaces the A725 which is a bit confusing, but is optimised for smaller size and lower power.
  
  And the C1 Nano replaces the A520 (latest in the same line as the good old A53) as a core that also exists and powers your budget tablet probably.

• Speaking of budget tablets, I have my Lenovo Idea Tab and Lenovo Legion Tab charging right now ready for testing tomorrow. (Notebook Check)
  
  And a couple of older models I dug out to see how things have improved over the years.
  
  Couple of things immediately evident:
  
  First, the more expensive Legion Tab boots much faster, which it should do since it has a much faster CPU (Arm X3 vs. A76 cores).
  
  Second, while the Legion Tab has a glossy screen, the Idea Tab is matte. Very matte. Lenovo's marketing material describes it as "paper-like" and at first glance that is correct.
  
  
  
• UTF-8 is a brilliant design. (I Am Vishnu)
  
  A brilliant implementation of a terrible idea.
  
  UTF-8 is the default implementation of Unicode, and Unicode is an attempt to create a single alphabet that can encode every human symbol ever, from all languages including ones we can't read and ones that have syllabaries or pictograms rather than alphabets, and also everything else.
  
  But it's just an alphabet. There is no embedded context as to what language you are using if the same symbol appears in more than one. Which happens all the time.
  
  Which makes it impossible to write anything that can be read unambiguously.
  
  
  
• AI-generated medical data can be used without ethical concerns, say exceptionally unethical people. (Nature)
  
  If you are wondering how the AI learned to generate useful fake medical data, then your most likely guess is precisely correct: It was trained on real human medical data.
  
  The issues here are large enough and obvious enough to provide roosting space for the entire former eastern seaboard population of the passenger pigeon.
  
  
  
• The employee of a DVD manufacturer who leaked the Spider-Man Blu-Ray has been sentenced to nearly five years in prison... (TorrentFreak)
  
  ... On unrelated firearms charges.
  
  
  
• Samsung is releasing the Galaxy Tab A11 range, consisting of unusable low-resolution garbage. (Liliputing)
  
  1340x800 on even an 8" screen is unacceptable. Google fixed that in the Nexus 7 in 2013.
  
  
  

posted by Pixy Misa at 04:30 AM

| Access Comments