Episode 407

407: It Isn’t a Big Grapefruit

Marco's Mini reviews, more on M1 scaling, completely predictable computer waffling, and our problem with Apple News.

December 3, 20201h 49m

Audio is streamed directly from the publisher (traffic.libsyn.com) as published in their RSS feed. Play Podcasts does not host this file. Rights-holders can request removal through the copyright & takedown page.

Original episode page

Show Notes

Pre-show: You heard it here first: Marco is really starting to justify a new Mac.
Follow-up:
- You can restore stuff from iCloud here (via Ryan Fegley)
- Accelerated TensorFlow for macOS probably just uses the GPU
  - Neural Engine can be used as well, though. (via @hishnash)
- Apple did make an ECC RAM controller in the past (via Max Lein)
- Lots of info about scaling the M1 via Jonathan Dietz
  - Die size spreadsheet
- Cerebras, the big-die ML accelerator (via Andrew Bunner)
- Memory bus thoughts from Wade Tregaskis
  - Memory interface bandwidth
- Dual M1 Macs
- “Low-effort” Mac Pro
- TMSC Achieves Breakthrough in 2nm Manufacturing
  - What is “risk production”?
- Alexander Graf virtualizes ARM Windows on Apple Silicon
  - 9to5Mac
  - QEMU
Marco’s Mini Monologues
- Smart Speaker Size Comparison (scroll down a bit)
- Panzarino’s observation about Qi chargers
- …and his FU
#askatp
- Is it possible to disable the fan on the new M1 MacBook Pro? (via Matthew Taylor)
  - CoolBook
  - TG Pro
- What’s our deal with Apple News+ anyway? (via Bryan Hoffman)
- Component Motherboards : SoC :: ICE : Electric Cars? (via Kevin Doran)
  - VLSI
  - VHF / UHF
Post-show:
- Yes, John is aware of breadboxes.
- Tesla’s “Mission Accomplished”