Stay out of the race sounds insane when AI is, and will be, a common commodity. China is getting great bang for their buck on AI training, why not the same for Europe?
The vast majority of people simply don't need SOTA models. I run most of my programming on a single 7900 XTX with Qwen 3.8 27b. I'm more interested in scaling out to run more instances of the model in parallel, than I am in scaling up to better models.
We're also seeing behemoths like Google and Meta both fail to keep up with OpenAI and Anthropic. Microsoft has given up on SOTA training. Loads of once promising LLM training companies are no longer relevant at the SOTA stage such as Cohere, Mistral.
I've said this countless times but SOTA LLM market looks like a classic monopoly/duopoly market over time. Each generation requires magnitudes more resources to train and you can only compete if you've made enough money on your previous generation.
On the contrary, small efficient models using recent improvements should be more sustainable than huge frontier models that require billions and years to train
The obvious draw-down on resources being used by companies following the token price hikes?
The fact that in the testimonials both frontier labs quote the same company (who I'm led to believe is still tokenmaxxing).
The fact that we've not seen actual productivity benefits or new products resulting from people outside the companies using it and a lot of things that in most businesses would result in a lot of investor skepticism and concern.
The fervent development and focus on 'the next big thing'.
All the mainline kernel support is reverse-engineered. No help from Arm.
Thus there is a considerable lag for the support of newer Mali generations.
Some Linux distributions for various embedded computers with Arm-based CPUs and GPUs use closed-source GPU drivers from Arm, not the reverse-engineered Linux drivers, when the Mali versions are newer than supported by the kernel and Mesa.
The Arm GPUs do not have public technical documentation, not even at the level of NVIDIA, much less something comparable with AMD and Intel.
Most Mali drivers from the kernel are not contributed by Arm, but you are right that nowadays some contributors from Arm have appeared for the older reverse-engineered GPU drivers, like Panfrost and Panthor, and now there are also a few GPU-related drivers, like HDLCD and MALI-DP, which have indeed been contributed by Arm.
Besides the Mali generations that are supported in the mainline kernel, there are many other Mali generations for which Arm has not provided anything else except closed-source drivers for Android, in a few cases also closed-source drivers for Linux. However, there are many other reverse-engineered Mali drivers, besides those from the mainline kernel.
It is good that Arm has begun to contribute to the kernel DRM drivers, but these drivers do little else except providing the means to communicate with the GPUs. The complex work is done in the user-space GPU drivers from Mesa.
The Mesa Lima driver, for older generations of Mali, does not have any contribution from Arm.
In the past, Mesa Panfrost was also completely reverse-engineered. Seeing that now Arm has begun to contribute to the kernel Panfrost DRM, I do not know whether they have started to contribute also to the Mesa Panfrost.
In any case, while some contributions to the previously reverse-engineered drivers are much better than nothing, what would have been much better is to publish the technical documentation of the Mali GPUs, in which case it would not have mattered whether Arm contributes something or not, as anyone could have written the drivers.
I liked it better when electronic components which didn't have very thorough detailed documentation were instantly recognizable as completely unsuitable for a production BOM. :\
>All the mainline kernel support is reverse-engineered. No help from Arm.
That's not true. Arm has been supporting the open source driver effort for the last couple years, but it's still behind the closed source one in terms of performance, features, and HW support.
The Pixel 10 and 11 GPUs are offensively bad. Not only are the Power VR GPUs slow, and less capable feature wise than Mali and Adreno (Xclipse is still around too!), but the tile Google put in the Pixel 11 is actually an _older_ design than what they put in the Pixel 10. They just clocked it higher to make up the performance difference.
And then Google has the gall to charge basically flagship prices for their under powered junk.
Personally I don't buy flagship phones because I don't use their full compute capabilities anyway, but a Pixel 11 would cost 2x what I paid for my Pixel 8. For 0 GPU performance improvement.
I have fond memories from back when Pixels were actually flagship. The Pixel 2 is my nostalgic pinnacle. I'd totally forgotten about that squeeze gesture until I got my old Pixel 2 out to erase.
I just can't wait to see it being tested in an actual phone. So far all the data are done on development board. If they are really as good as everything we have read so far, both Mediatek and Qualcomm are in big trouble.
I think tech enthusiasts like thinkpads for the same reason they dislike macbooks: out of the box linux support, good ergonomics and repairability.
Do thinkpad owners care about battery life? Yes, but once you are past a full work day battery and can charge it with same USB C charger that came with your phone it becomes a non issue.
Lenovo and Linux support should carry an asterisk. Hopefully their acquisition of Phoenix fixes much of the "too quick to market" problems they seem to keep having.
My ThinkPad Z13 (aka AMD Carbon) was a piece of garbage: arrived without Fedora fully installed, crashed at least once per week, had GPU and dock and USB and sleep problems, sent back for repairs multiple times. I tried to return it after the multiple failed repairs and they refused. My greatest joy was that insurance covered it being stolen from my apartment in Dublin.
My X1 Carbon Gen 11? Not great. Crashed 7 times in the 9 days I had it. Didn't sleep properly. Ibus and wireplumber would randomly crash. I'd get a glib error if I type an invalid Bash command. I tried Fedora 38 and Kinoite, Ubuntu 22, Debian 12, RHEL 9.2, TUXEDO, Pop, and Tumbleweed. Nobody had the right mix. Red Hat didn't even make it through the installer. Debian overheated rather than suspended when you close the lid. I learned about and trialed all sorts of kernel flags: intel_idle.max_cstate, iommu=pt, Intel_iommu=on, Intel_iommu=pt, usbcore.autosuspend=-1, i915.enable_dc=0, i915.enable_psr=0, acpi.ec_no_wakeup=1, thinkpad_acpi.fan_control=1, acpi_osi, and a few others. Sometimes the 10-second power button shutdown was ignored; you have to use the little reset hole.
I can't even fully blame Lenovo. Their Linux lead puts in a ton of effort to do everything right. BIOS seems to be a major weak point. I'll read BIOS release notes for the laptops I hated, and they'll have a list with e.g. "fixed fingerprint reader problem, fixed hanging after hibernation, fixed hanging after resuming with lid closed, wake now works with any keyboard key instead of just F-keys, fixed issue where computer restarts unexpectedly, fixed overheating while charging, fixed problems related to docking and external monitors, fix sound, fix backlight, fix power profiles, fix slow boot, fixed USB port refusing to charge or enumerate (rarely (yes they actually said rarely in the release notes, this problem affected at least me)), improve lid behavior, fix touchpad ACPI, fix sleep from power button, trackpoint doesn't work if lid closed with suspend disabled (fixed), touchpad stops functioning after restart (fixed), system hangs during Modern Standby (fixed, we hope?), bunch of CVEs, camera shouldn't suck anymore, updated copyright year" This is less than half of the changes and was paraphrased from the actual Z13 release notes.
I mean... G. F-ing D. With so many EC/BIOS problems on nearly every peripheral, I'm baffled that this got released in the first place.
Also, if you're going to sell Linux laptops, at least have Linux support at the service center. On multiple occasions, I got a call, "we ran a Windows USB image, we didn't find any suspend problem."
My X1 Carbon Gen 12? Acceptable. Gets hot easily. Every now and then (twice a month) just stalls, sometimes I can unplug and close the lid and come back later and the OOM killer will finally do its job. I'd put 90 percent of the blame on KDE Plasma and Fedora. I almost blame myself for having too much stuff open at once, but then I recall that "having a lot of work open" generally can be done on older laptops, well-built desktops, and Windows machines.
My X1 Carbon Gen 13? Did not NOT have problems until Ubuntu 26.04. Screen would glitch green, some hardware was unsupported---in particular, wireless---had suspend problems, overly long boot, one CPU would always run at max (gnome-shell at 100% of CPU 0, whole machine at 13%). The biggest problem was learning that full-disk encryption and secure boot don't play nicely with installing the HWE kernel (which I would've needed for the unsupported hardware), requiring a full backup and reinstall while I was trying to start a new project during my probationary period at work. I need Ubuntu for work, and my colleagues were already annoyed that they couldn't find a Legion 5 with ANSI keyboard from a German vendor so that we would all have a common platform. Unfortunately, Rust coreutils and GNOME are still buggy in 26.04 and it affects my workflow, but at least the hardware (finally) works.
Partner's P-series? No problems so far. Coworkers' Legions? no complaints from them. Old-ass X301? Works fine. T480? Nice machine.
Lenovo itself has horrible dark patterns on their website (defaulting to 3-year premium warranty in the EU despite this being an optional upgrade and restoring this default every time you modify the device config from the checkout cart, market-specific pricing and customizations despite all units originating from the same factory such that Linux from another country was a better deal than "No OS" from my own country, basing price decisions on how long you view specific pages). The fact that the website has many glaring bugs (including in the feedback forms to report bugs) is not a great look for a computing company.
My dream machine was the XPS 9370 Developer. Dell went the wrong way for me with the bathroom shower tile keyboard and mostly OLED screens and no more explicit Linux support. If anyone can capture that lightning in a bottle again, I'll buy four.
> same USB C charger that came with your phone
Yeah, you'll need that phone charger for your Lenovo, because the latest few generations of Lenovo chargers break in a way that they'll only charge in one orientation for a while, and then they eventually break completely. Luckily, they do this quickly, so you can replace them under warranty for as long as you're willing to tolerate the issue.
-----
I choose Lenovo because every other PC laptop (plus quite a few Macs) is/are worse, but I can't call ThinkPads "good"
I dont understand why EU is not stepping in. LG literally spies on you by capturing screenshot of whatever you watch and send them to the mothership. This includes images and videos from your private USB drive.
The surveillance state is not stepping in to stop spying? This is stuff they can use to profile you. The EU Commission is very keen on Chat Control and has pushed a lot of surveillance laws.
Governments only protect privacy when there is a lot of pressure to do so.
I belive you are mistaken. Android 17 final was "released" 6-7 weeks ago.
But Google being Google, this release is pretty much useless until Samsung et al deal with all bugs and performance issues, which will take 2-4 months.
As a Linux, macOS and Windows user it’s really strange to me that you can’t just install the latest OS on your device. Even my iPhone updates to the latest version as soon as it’s released.
Seriously, why can't Android just be installed? Are they building it like the old-timey kernel before modules and embedding drivers in a giant monolith or something ridiculous?
Except there were websites requiring you to install ActiveX and proprietary plugins to do anything meaningful as soon as 1995? Also banks requiring proprietary security modules to access your account, they were pretty ubiquitous in the early 2000s where I lived.
The difference now is that accessing services through the internet is a necessity rather than a luxury. I guess passing laws could fix this, but it's exceptionally hard to get anything done through the barrage of big tech lobbying.
(warning dumb question incoming): since android apps are java, is it possible for a third contender to create a jvm for their platform and some hooks for android api’s and run anything that was made for them?
Yes. The main issue is hardware attestation, e.g. the bank's backend refusing to allow you to log in unless you can prove cryptographically that you are running a Google-certified device and operating system.
Doesnt look like it's still maintained, but it shows it is possible.
I completely forgot about waydroid due to Wayland not being ready for general use.
(Aside from accessibility stuff mentioned on hn numerous times, still doesn't allow keepass to type passwords last I checked. Admittedly, I check Wayland about as often as I check Hurd.)
I almost never use dedicated apps. I am able to access all my bank and other financial accounts through their websites on mobile, no problem. (Vanadium browser on GrapheneOS)
reply