I tried creating an emulator for CPU that is very well known but lacks working open source emulators.
Claude, Codex and Gemini were very good at starting something that looked great but all failed to reach a working product. They all ended up in a loop where fixing one issues caused something else to break and could never get out of it.
When they get stuck, I find adding debug that the model can access helps. + Sometimes you need to add something into the prompt to tell it to avoid some approach at a point.
I’ve been trying to do the same thing as a hobby project to just imagine some “what ifs” with some slight changes to the original 8086 and the 80286.
It just never produces an actually working result without a lot of intervention on my part. (My change was merely changing the paragraph size from 4 bits to 8 bits on the 8086.)
Interesting. When I had Claude write a language transpiler it always checked that tests passed before declaring a feature ready for PR. There was never a case where it gave up on achieving that goal.
Please tell me what CPU it is. I would give it a try. I doubt strongly a very well documented CPU can't be emulated by writing the code with modern AIs.
I dont think Google is gimping your device to sell you new ones.
What you probably see is Google toggling on some new AI feature, which is now doing some initial on-device computation. It will usually calm down after a few days.
> I dont think Google is gimping your device to sell you new ones.
I'm not sure.
Last year they ruined my Pixel 4a by pushing that battery patch to everyone (even to end of life devices).
Their official repair center replaced the battery as per Google's guidelines and offer but during the battery swap they managed to physically break other parts of the phone over the course of multiple visits. Each visit broke a new component.
Google support didn't do anything in the end, eventually ignoring me. This went on for almost 6 months with an email chain over 100+ replies.
Eventually the repair center gave me enough credit towards a new phone (this 9a) but only after I mentioned I was going to small claims court since they left me with a device that didn't function in the way it used to function before I visited them.
There are tons of evidence showing that cameras are alone are not safe enough and even Tesla has realized that removing lidar to save cost was a mistake.
In raw power x64 is still the king, has always been.
You can skew things by looking only at single core performance (where apples most expensive cpus might win because of their strategy of having fewer but more powerful cores + memory latency gains are much more visible with only one core).
With that said, things are changing in the PC landscape and some extremely powerful and power efficient ARM designs are coming soon. We have already seen a small glimpse of that with MediaTek.
I just looked into that for this thread, but the highest-end Apple chips (M4 Max/M5) outperform the fastest normal Ryzens in both single core and multi core tests.
They're awful for gaming, which most benchmarks online are run for, but it takes a Threadripper to dethrone Apple's impressive CPU performance.
Of course, this also has to do with their integrated memory architecture, which isn't very popular with high-end PC customers who like the ability to upgrade their RAM.
It is hard to compare apples to oranges, sometimes people benchmark different software (e.g. Safari on osx vs Edge on windows) or software that on one platform is more optimised. But in general it looks like AMD despite using an older process node is doing fairly good.
Which probably doesn't help much you if you dont know about XSS.
reply