tedd4u
Joined 17 years ago
Karma
1,707
Submissions
555
Member since
Aug 17
Recent submissions
Comment · 2 days ago
One question is how do you get to the context menu with touch. Tap and hold?
Comment · 3 days ago
100%, all these product changes are so crazymaking.
Comment · 3 days ago
They have it. It’s called Xprotect. I’m surprised it hasn’t been used to remove the bad certs.
https://fullmetalmac.com/cybersecurity/macos-security/xprote...
Comment · 3 days ago
True. I still think it is good. My dream is for 120Hz on a future Air model. I took it as a positive sign they brought 120Hz to the non-Pro iPhones with the models last year. I can dream, I guess. And yes, I know MBP screen has benefits other than the refresh rate.
Comment · 3 days ago
If that continues to be the case, next time you can "downgrade" to the MacBook Air. Even today you can get one with 32GB RAM, so when you're in the market I'm sure more will be available. They are quite thin and light compared to the MBPs.
Comment · 3 days ago
There's a whole movie about Golden Age thinking / Golden Age fallacy by Woody Allen called "Midnight in Paris." Worth a watch.
Comment · 3 days ago
And the iPhone XR! All part of the twelfth-gen models. Of course the subsequent thirteenth-gen models were just called "iPhone 11" in a return to sanity.
Comment · 3 days ago
Yeah I think this is probably right. I use a 13" MBA as my primary machine so space is crucial and Liquid Glass wastes it.
Comment · 4 days ago
Isn't the laptop version of the 5090 is a more apples-to-apples comparison? It has 896 GB/s memory bandwidth (compared to the 614 on the M5 Max MBP).
Comment · 4 days ago
Liquid Glass is more than just the transparency. There are many other changes that cannot be disabled that way. Not to mention the overall lag compared to Sequoia (I skipped Tahoe entirely so I had a direct knife-edge transition). 27 is less bad than 26 but Liquid Glass doesn't make anything better than prior UI and makes several things worse, even if you turn down the transparency.
Comment · 5 days ago
OpenAI itself cannot answer these questions. A prospective price-sensitive (or speed-sensitive) user has to experiment to find the cheapest (or fastest) model that is sufficient for each task.
Comment · 5 days ago
An example is I like Claude a lot, but for the same $20 spent with Google to get Argon 4, I also get ad-free on YouTube, a ton more storage, and it can access all my Google Drive files natively. I’m not going to pay for both. And I already stopped paying ChatGPT last Nov when Claude got good.
Comment · 8 days ago
Not to mention raise your electricity bill, drain your aquifer, make loud noise constantly, and belch diesel exhaust and get even louder as the diesel backup gennies are periodically tested. Maybe you're extra lucky and it has ongoing / indefinite diesel reciprocating or even jet turbine behind-the-meter generation!
By the way, your municipality may have drawn interest by excusing the builder/owner from most property taxes, too, so it won't support your schools and essential services.
Comment · 9 days ago
“Claude, please port everything over from Dots to Duet. Make no mistakes.”
Comment · 12 days ago
There are a huge number of citations about looping on this Ouro page [1]. Ouro is an open looped LLM with a very instructive paper at [2] (also containing references to lots of helpful papers). Basically they get these four results from looping:
• Parameter efficiency. 1.4B and 2.6B parameter looping Ouro models match 4B and 8B transformers on most benchmarks.
• Adaptive computation. Looping is trained in a novel way to avoid exiting too early or too late.
• Reasons to use looping. Looping does not increase raw knowledge storage (approximately 2 bits per parameter for looped and non-looped models) but dramatically enhances knowledge manipulation capabilities.
• Improved safety and adherence. They find that safety and alignment of CoT to final output improves as recurrent steps increase.