This is a short term loss for a potential long term improvement. By eliminating dependency on translation APIs they can force the use of more open solutions like oneAPI which is even getting buy-in from companies like Imagination.
Keeping cuda alive is a bad idea.
The bot avoids roasting torvalds but will roast maintainers. That’s a little odd, but I guess it keeps it out of the news.
TFIDF and some light rules should work well and be significantly faster.
Yo, they added full page copies now? Gotta give it a spin again
Everything loops back to steam in the end. Solid state thermoelectric devices have been around forever, and before that we had the idea of using thermal energy to augment magnetic fields and jump to kinetic energy without any intermediary conversion. All very low yield results, but we’ve tried it anyway.
Keep thinking about it, we need all the brains we can get, but don’t write it off as a novel idea that the other egg heads just haven’t gotten around to solving yet.
Data centers will probably be the only practical application. Consumer electronics will probably barely produce enough energy to power the regulator and tie-in circuit just to feed back into the pwm driver for fans nowadays.
Some internet funeral aesthetics
Lol, changing the country of origin doesn’t constitute innovation from a consumer standpoint…
Now if this was using 5nm or chiplit or any of the other buzzwords of the day it could be marketed as innovative in the modern sense of the word.
Realistically there is no innovation left for ARM platforms. They all use the same core schematics. They only control data flow and peripheral IP as a manufacturer, unless they feel like building their own core from the spec (nobody really does that anymore as ARM has been desperately trying to standardize everything). The most “innovation” I’ve seen has come from stubbornness around keeping legacy bus architecture around instead of adopting AXI (even when all the IP they are trying to use already uses AXI and they keep having to make translation hardware).
Fuck it. Gun it at the brick wall. Jerry’s rigging up an emergency break as we speak. Don’t mind that the last piece to said break may be missing.
- Man who will probably die before we hit the wall
I was working at a company at one point that got a contract to build something I viewed equivalent to malware. Immediately I brought it up to several higher-ups that this was not something I was willing to do. One of them brought up the argument “If we don’t do it someone else will.”
This mentality scares the shit out of me, but it explains a lot of horrible things in the industry.
Believing in that mentality is worse than the reality of the situation. At least if you say no there’s a chance it doesn’t happen or it gets passed to someone worse than you. If you say yes then not only are you complicit, you are actively enforcing that gloomy mentality for other engineers. Just say no.
Publicly traded corps just saw the letter “p”, assumed profit, and announced a 5 year plan to discover the rest of that sentence. Their shareholders are still upset it took them a whole letter before they had a plan…
To me on the security side of things caddy has a feature I have yet to see anywhere else: default reverse proxy headers.
Got something you want to lock down remote js loading on unless it explicitly requests an override? Default the variable to a locked value. The application can override it with it’s own header as necessary.
deleted by creator
Search telemetry in a web browser is absolutely insane. I can understand more usage statistics but search telemetry just makes it sound like they want data on who to make an offer to for the next default search provider slot.
Or worse yet, another half-assed partnership with some sketchy 3rd party with a completely fucked moral compass and a privacy policy to match.
Ha, gotta add this regex to my spam filters now that you’ve pointed it out
Good video going over practical pros and cons currently:
Googles requirements for ARM cores on Android was pretty high. Don’t think I’ve seen a RISC-V core get close yet…
I wish nginx had the concept of default header values for reverse proxies…
I mean, you can kind of do it with macros but man…
Because ARM was built to be cheap.
BIOS nowadays is basically a bootloader shim in EEPROM. The majority of the ARM ecosystem wanted flexible and cheap devices. This promoted the use of a small ROM loader burned into the device and a removal of basically all EEPROM from the SoC.
The flexibility came back through the use of a secondary bootloader layer normally stored in the devices primary storage. Most manufacturers use u-boot or coreboot on an SD card or eMMC. Android standardized this as part of their partitioning scheme. All devices have a dedicated bootloader partition housing the secondary bootloader and any additional boot artifacts.
Then phones became wildly expensive and invalidated most of this.
It already has. Most of what ARM is doing to be cheap was already pioneered by PowerPC.
ARM EBBR specifications attempt to standardize this boot flow somewhat, introducing a standard EFI shell in u-boot. This does not solve the dependency on the secondary bootloader, and it doesn’t prevent people from shooting themselves in the foot. It just makes distro interactions with the secondary bootloader more standardized.