saidinesh5

Back then everyone believed that. Even Android vendors. Motorola Droid did sell well. Touch screen typing was really annoying until Swype really took off.

Their Qt story was mostly geared towards their Linux phones. N900, N9 were the last of the kind. They killed the N9 before even it was launched. (The infamous Burning platform memo).

At that time, I genuinely believed that the N9 had the vision and design to beat anything Android was - at that time. Till date, it's still the best UX i experienced in a phone. Even a lot of Nokia managers I talked to called Elop a trojan horse from Microsoft.

But given how Android's source was open while Meego was anything but.. along with the whole Nokia + Microsoft drama... It's just a shame

> Hardware accelerated video decoding (for video playback or for WebRTC) is available via VA-API for both X11 and Wayland. Since firefox-101.0.1-4 package it's enabled by default for Intel/AMD users.

From https://fedoraproject.org/wiki/Firefox_Hardware_acceleration

I think that was around 2022-23.

I remember this affecting my laptop video watching a lot. Google meet feeling sluggish on Firefox is also something I remember until very recently..

At my current job, one of my biggest blunders was thinking "Let me just order the same hardware as most of my teammates to avoid unnecessary complications".

Most, if not all, of our current work happens on remote cloud vms. Now I'm stuck with carrying a 3KG monstrosity to work.Every day.

Absolutely no positives compared to my ThinkPad that weighed less than half in my previous job.

Remote development is "good enough" these days. With VS Code, Development containers etc... Having a light weight, portable laptop is so much nicer than a laptop that you can't even rest on your lap for long duration.

The only thing you need to be mindful of using light weight laptops is not having enough RAM to fit all your browser tabs..

Lately I genuinely believe that the future will be large frontier models generating and updating inputs/skills for "good enough" local models to solve our daily problems.

A lot of tasks which need a bit of intelligence don't really need that much compute. Just good enough documentation / skills, tool calling and a good enough local model.

Not sure what exactly this means for all those data centers that are getting built... But exciting times.

This is actually an interesting idea...

Especially if it gives us an insight into what prompts were given to generate something we approve of vs. something we disapprove.

Not just unedited but what feels like low information density...

For eg. If you're generating a whole document with a single prompt, say "write a document on how to implement this"

vs.

"Find various approaches to implement this. These are the approaches we already have. These are the constraints we currently have."

Then verify the generated text, remove useless hallucinations. Use the same AI to verify the generated output.

You're basically out of luck. This is happening all over the world these days with various government services.

It's not just for clients with no smart phones. It's also clients who activate developer mode on their smartphone etc... clients who use VPN. Clients who run AdBlocker. Clients who use Firefox.

There needs to be a severe push back against this.

Things have gotten so crazy that the homeless here are walking around with QR codes printed out when begging for money. I'm guessing they're a network of people who share the same smartphone/payment method..

A lot of "non Android Linux distributions" like sailfish os actually use a lot of Android infrastructure to keep working.

Drivers doesn't just mean the kernel. It's the user space binary blobs and services that need to talk to the kernel to enable the hardware.

Other than that there's waydroid, alien dalvik etc.. that run another Android instance in a container.

The thing is a lot of Android applications use safety net/other methods to make sure they only run on. "Approved"/stock hardware.

The bank website typically needs the app to enable two factor authentication in a lot of places.

For eg. There's no browser based alternative to make UPI payments that i know of.

Has anyone major used that for any production project though?

> What's wrong with pixel drivers for instance? I'm of course not talking about those written by e-waste ODMs like Tinno who can't even release GPL kernel sources properly

A couple of issues , back from my sailfish os porting days (haven't touched it in half a decade so things might have changed):

1. You would still need to run a lot of Android user space to make use of these drivers. We had a separate second init system that runs a stripped down version of Android init, starting enough services to make use of these drivers.

2. These binary drivers are tied to the kernel and Android versions that were released with the phone. So if you wanted some of the newer kernel features (back then it was like better container support), you're out of luck.

Usually, all this means you rely on a lot of hacks to fix some bugs... And porting takes a lot of time and effort.

I tested out last year's foldable phone for a couple of months.. borrowed it from a friend.

Maybe I'm just losing interest in phone tech - but to me, it was only down sides and no upside.

The phone was bulky. Fragile.

The larger screen you get when you open the phone - was absolutely useless for most of the content I watch. The screen is 4:3 and most of the content I watched was 16:9. The black bars at the top and bottom of the video meant that the actual video content is exactly the same size as my non foldable Motorola Edge 40.

I tested out Pixel 3 (or 4 .. i don't remember at this point). The 30 hours is when you have all the functionality enabled all the time. Like the always on display, always on notifications etc ...

I used to auto set the watch into DnD / battery saving mode at night, and used the "raise hand to turn on the display" functionality.. I used to get 3-4 days of continuous use.

Charging used to be like "put the watch into charger after waking up. It'll be ready by the time you get ready for office".

I did miss the 30 day battery life of my Amazfit bip and 2 week battery life of Pebble before that.. but charging once a week wasn't all that bad...

But again, I mostly used my Pixel watch as a fitness tracker. Didn't really care for most of the apps - except maps.

I think that's just the phase 1? Once we have all the use cases figured out, optimizations gradually follow?

I intend to.

They don't sell steam deck/it's parts where I live. So the next time people visit me from far away, they'd have to bring one along.

That being said, steam deck isn't as powerful to play some of the games like Spider-Man 2. The stuttering is quite high on that.

This is a really good card that keeps on getting better.

After my steam deck battery puffed up and my RX580 became too old to play any modern games, I had two choices:

1. Upgrade to a new GPU - this was costing me $300 at the very least for a 3 year old card

2. Hope to get a steam machine - but the crazy queue and the $1000 price tag really soured me on this

3. Get a bc-250 for $150 and make do for a year or two before upgrading.

I'm so glad I went with this option. I just used my old atx power supply and old PC case. Everything almost just works. It consumes lot less power than my previous gaming PC. It's a lot more silent. Some times there's a crash or two. I'm waiting to upgrade the power connector before unlocking it's full potential (40 compute units, and the two extra cores). So far I've finished playing Spider-Man 2, Hogwarts legacy, now playing 007 first light.

I think the steam deck proved otherwise too..

I haven't had enough motivation to sit on my couch and game after a long day ..

But the same game, in bed, on my deck was so much nicer..

All I can now say is having a dedicated device, that's not your laptop/computer to play games is definitely a market - be it Steam machine (/custom builds), hand held gaming, or just regular consoles..

Chances are you'll run out of ram trying to run a browser inside another browser on a TV with 1-2GB ram usually.

You might have a better luck with network level ad blocker like pi-hole or adguard for eg

Wait, what happened to quick share?

It's definitely the port too. I changed the cable and the connection is still loose. Not as bad as the previous cable, but still bad..

Yesterday, I had an interesting situation where a colleague was using LLM to type up the chat responses to me.

They used an LLM to come up with a Hackathon idea. They tried to use LLM to get to the implementation part. One of the permissions they (/the LLM) were asking for wasn't even needed for their project. When I pointed that out, I got a response from them that starts with "You are right about the permissions".

Today they've successfully finished the project too.

Not sure what I even think/feel about this but it's definitely weird to think about this.

I think my only big complaint with USB C these days is that the sockets wear out quickly.

The charging port on my motorola phone got so loose that I frequently end up with an uncharged phone after a full night of "charging", just because the cable's own weight keeps yanking it out of the phone. The phone is not even 3 years old.

It would be nice to get a user replaceable USB port. Meanwhile, I'm not sure if those magnetic USB cables help this situation any bit or they further damage the port.

The best part of this is you can daisy chain tunnels and have a network connection between any machines you can ssh into.

If you have a private server with a public IP, you don't need localtunnel etc.. you can just use ssh tunnels to expose your home network services over public IP.

> I don't quite understand why "embedded" has historically meant learning a completely separate tool. I've been building for x86 and ARM servers for 5+ years now, so why should targeting an ARM board be any different?

Because embedded usually means dealing with complete functionality provided by the soc, bare minimum for the init system to get quick boot times.

If you just want to boot the CPU and run some IoT app, device tree + kernel + uboot should be enough.

But for anything with a GUI, You'd probably need the vendor provided binary blobs for GPU accelerated UI, proper gstreamer packages with all the patches to make it work with the binary blobs... properly packaged Qt etc ... Not to mention read only rootfs and A/B partitions for upgrades to deal with power loss headaches.

I haven't used Nix for any of these but most of these things are available ready to use with the vendor provided buildroot or yocto setups. And it would've been just more work to get this all working with yet another build system/package manager

$1049 .. Damn.

Here's hoping my $135 BC-250 arriving tomorrow works without any issue.

Either way, congrats to Valve!

Wouldn't HEVC licences already be paid by the hardware/gpu vendors on most devices? And Microsoft just exposes api for that hardware?

Is this just for a purely software implementation of it?

Fun story: A few years ago, I worked for a small company that customized off the shelf routers to enable businesses provide Wifi Hotspots.

The routers were very basic model with very limited flash memory (~4MB?). I was brought in to build firmware for those routers. I ended up customising openwrt - removed all kinds of packages to make their packages fit on those routers. At the end, I had less than 4KB space, And I needed to implement a "heart beat" service. A lot of routers were behind firewalls that only allowed http, https and a couple of other protocols. Libcurl was too heavy. So I ended up writing a shell script that used this feature of bash to send out heart beats.

Fun times...

We faced a lot of this this year. Eager new joinees sent 100s of lines of CLs to review to "improve" little things.

Not only did the new changes did not fix what they thought it would fix but it broke other things in unexpected ways.

I brought in two changes after that:

* I'm not reviewing/reading anything that you yourself did not read / test in the target environments properly. If all it takes is an LLM prompt, I could be issuing the same prompt to make my life easier.. and If you're sending a CL, you should be owning the code you send.

* Me being more involved in the design process so review burden itself becomes lower. A bit of pair programming from time to time helped too.

Not sure how things will turn out after this but so far they seem better.

Typically libgcc_so.so is loaded by the linker, which uses an mmap call to map the binary into the address space.

> The kernel keeps track of which file is mapped where, and can detect when a request is made to map an already mapped file again, avoiding physical memory allocation if possible.

Relevant stack overflow answer: https://stackoverflow.com/questions/61950951/linux-shared-li...