CalcProgrammer1

A Framework phone with 2 modular Framework sockets would be amazing. I don't care if it's thick. Make it repairable and support Linux Phone OSes like postmarketOS and I would absolutely buy it.

Squeekboard is where it's at. By far my favorite onscreen keyboard for Linux and mainly because you can easily create your own layouts using .yaml files. I'm tired of virtual keyboards that omit keys needed for development and terminal use or shove them off to separate tabs. My custom Squeekboard layout fits my needs exactly and I'm pretty fast at typing on it (typing this on it now). I wish it were usable outside of Phosh, though tbf I haven't tried. Between GNOME Mobile, KDE Plasma Mobile, and Phosh (Squeekboard), I choose Phosh primarily because of how much I like Squeekboard.

Except that in the case of VGA (and DVI, HDMI, and DisplayPort) the i2c interface is intended for use over the cable. All of those ports have a pair of i2c pins and corresponding wires in their cables. The i2c interface is used for DDC/EDID which is how the computer can identify the capabilities and specifications of the attached display. DDC even provides some rarely-used control functionality. Probably the most useful of which is being able to control the brightness of the display from software. I use the ddcci module on Linux and it lets me control my desktop monitor brightness the same way a laptop would, which is great. I have no idea why this isn't widely used.

Edit:

This i2c interface is widely used to control the lighting on modern graphics cards that have RGB lighting. We've spent a lot of time reverse engineering these chips and their i2c protocols for OpenRGB. GPU chips usually have more i2c buses than the cards have display connectors, so the RGB chip is wired to one of the unused buses. I think AMD GPUs tend to have 8 separate i2c buses but most cards only use 4 or 5 of them for display connectors. There is also an i2c interface present on RAM slots normally used for reading the SPD chip that stores RAM module specifications, timings, etc. This interface is also used for RAM modules with controllable RGB lighting.

Yeah, the lack of proper discoverability on i2c truly sucks. You have to just poke random addresses and hope for the best to see if an i2c device exists on the bus. It's a great standard but I wish it would get updated with some sort of plug and play autodetection feature. Standardized device PID/VID system like USB and PCI would be acceptable or a standardized register that returns a part string. Anything other than blindly poking registers and hoping you're not accidentally overvolting the CPU or whatever because the register on your expected device overlaps with the overvolt the CPU register on the same address of a different device.

I mean, the open source driver already is out. The nouveau driver has been in the kernel for like a decade now. The userspace part has been in Mesa for just as long, though largely was unused due to nouveau not being able to use high clock speeds. That isn't the case anymore, and since the beginning of the year you've been able to test drive the new NVK Vulkan driver on nouveau with GSP enabled to get actually reasonable performance in several select games. NVIDIA isn't creating a new driver, they're contributing to one that already exists. Since this particular patch set is so huge I don't know it will make it into the next kernel release right away but this guy was the former nouveau maintainer, I expect he knows the necessary standards to get his code accepted.

Just fitting objects into the smallest box isn't everything according to the article. This is trying to identify fragile objects and recommend appropriate protective packaging where required to minimize the risk of damage in shipping. If you use a conventional packing algorithm to pack dishes and vases into the smallest box you will receive a box of glass shards on your doorstep. Is AI the best solution? I'm not sure, but using actual statistics of damaged goods and their means of packaging sounds like a worthwhile consideration.

This...actually seems like a good use of AI? I generally think AI is being shoehorned into a lot of use cases where it doesn't belong but this seems like a proper place to use it. It's serving a specific and defined purpose rather than trying to handle unfiltered customer input or do overly generic tasks,

I don't really see why they would hire him to achieve this goal. He had already quit as maintainer. He was out of the picture unless he resigned specifically due to accepting an offer from NVIDIA, but if that was the case and they wanted Nouveau stopped then why is he now contributing a huge patchset? If they hired him and he quit nouveau they could've had him work on the proprietary driver or their own open out of tree kernel driver, but they specifically had him (or at least allowed him) to keep working on nouveau.

Also, if they really wanted to EEE nouveau into oblivion, they would need to get every single prominent nouveau, nova, and NVK developer on payroll simultaneously before they silence them all because once one gets silenced why would any of the others even consider an NVIDIA offer? Especially those already employed at Red Hat? It doesn't really make sense to me as an EEE tactic.

What has been apparent over the past few years is that NVIDIA seems to be relaxing their iron grip on their hardware. They were the only ones who could enable reclocking in such a way that it would be available to a theoretical open source driver and they did exactly that. They moved the functionality they wanted to keep hidden into firmware. They had to have known that doing this would enable nouveau to use it too.

Also, they're hopping on this bandwagon now that NVK is showing promise of being a truly viable gaming and general purpose use driver. Looking at the AMD side of things, they did the same thing back when they first started supporting Mesa directly. They released some documentation, let the community get a minimally viable driver working, and then poured official resources into making it better. I believe the same situation happened with the Freedreno driver, with Qualcomm eventually contributing patches officially. ARM also announced their support of the Panfrost driver for non-Android Linux use cases only after it had been functionally viable for some time. Maybe it's a case of "if you can't beat them, join them" but we've seen companies eventually start helping out on open drivers only after dragging their feet for years several times before.

I'm cautiously optimistic. While I could see NVIDIA hiring him to stifle nouveau development, it doesn't really seem worth it when he already quit as maintainer and Red Hat is already working on nova, a replacement for nouveau. I got into Linux with Ubuntu 6.06 and remember the situation then. NVIDIA and ATI both had proprietary drivers and little open source support, at least for their most recent chipsets of the time. I was planning on building a new PC and going with an NVIDIA card because ATI's drivers were the hottest of garbage and I had a dreadful experience going from a GeForce 4 MX420 to a Radeon X1600Pro. However, when AMD acquired ATI they released a bunch of documentation. They didn't immediately start paying people to write FOSS Radeon drivers, but the community (including third party commercial contributors) started writing drivers from these documents. Radeon support quickly got way better. Only after there was a good foundation in place do I remember seeing news about official AMD funded contributors to the Mesa drivers. I hope that's what we're now seeing with NVIDIA. They released "documentation" in the form of their open kernel modules for their proprietary userspace as well as reworking features into GSP to make them easier to access, and now that the community supported driver is maturing the see it viable enough to directly contribute to.

I think the same may have happened with Freedreno and Panfrost projects too.

This is my cautious optimism here. I hope they follow this path like the others and not use this to stifle the nouveau project. Besides, stifling one nouveau dev would mean no other nouveau/nova/mesa devs would accept future offers from them. They can't shut down the open driver at this point, and the GSP changes seem like they purposely enabled this work to begin with. They could've just kept the firmware locked down and nouveau would've stayed essentially dead indefinitely.

Get some HDMI to VGA adapters, the kind that screw into the VGA port and then have an HDMI port. I have a bunch of old VGA monitors I use with Raspberry Pis and as test displays when working on PCs and never have to deal with the annoyances of VGA since they're basically HDMI displays now.

Most of the laptops I've seen the external port is connected to the dGPU.

Most gaming laptops these days don't support true GPU switching as it requires a hardware mux to switch the display between the GPUs. Every gaming laptop I've used from the past decade has been muxless and only used render offloading.