A lot of it just comes down to KISS and avoiding unnecessary overhead and indirection (like vtables, C++ STL containers etc.) so you're getting the most out of the hardware.
It's genuinely possible to do informative, fun and engaging videos on a niche OS; it does require a lot of skill and effort to pull off he does not have. His content has been way too shallow. There's a lot more to an OS than just the Desktop Environment being different, and 'Can it run Minecraft' (with missing/broken drivers he doesn't understand).
I've watched his previous videos on Haiku and haven't been impressed with him. He doesn't seem to have in-depth knowledge of how software and drivers work, which is especially important given how work-in-progress it all is, and that the lineage of Haiku goes back to BeOS. The community has done a lot of good work, and he doesn't grasp it fully.
Makes sense for them to support this given their scale. I wouldn't be surprised if GrapheneOS secures another big Android OEM vendor (after Motorola Mobility) and the user base of the OS quickly grows from its current 400k+ to millions.
Yeah there's no good alternative Version Control right now that can handle large binaries. And for Lore from Epic Games, while it's still 2-3 years away from being suitable for Game productions to replace Perforce, there's genuine interest and excitement about using/testing it (in some cases, studios are spending tens of thousands on P4 licenses they don't really want).
Lore is already used at scale in UEFN across a huge user base so the core tech behind it is solid and trusted.
I'm willing to give Tangled a go too with a project, but feature set to bridge the gap still has a long way to go (no idea how long it'll take). Github outages (especially when just viewing repos!) are getting way too disrupting.
I tried reviving my Asus Eee PC 1015PEM (1.5GHz dual core, 2GB RAM) and even running Linux Mint was a bit much for it (basic tasks were slow, and Wine had too much overhead for games). I initially tried upgrading the Windows 7 Starter to Windows 10, and while the upgrade did work...it was failing to even log in! Whole thing just seemed stalled.
Great article and fantastic work, though the commitment to stick with the ancient DCC software of the era (genuinely admirable), is an immense amount of pain, for no real gain that I wouldn't dare to try. Modern Blender can be extended to import/export legacy formats like DirectX 8 .x and N64 Homebrew for example have great asset tools all working in modern Blender, with a big community behind them. FBX was originally a (Kaldera) MotionBuilder format only back in that era and 3ds Max support for it didn't come until around version 8/9.
I'd say there's a way to build a modern asset pipeline for the OG Xbox where you can leverage USD (well supported in all modern DCC), and then build tooling/viewers around it to process/split/export everything out to DirectX 8 .x in a much more manageable way. Makes artist collaboration a lot easier (modern tools with guardrails inplace), and can measure and keep metrics on everything (every KB counts with 64MB RAM).
All these old versions of Maya, 3ds Max and Visual Studio are just a total liability now. Licences for them can't be activated, and even if they can, it's a lot of backflips to get Autodesk to help. 3dsMax/Maya 8 and earlier barely work on Windows 10/11 now (Vista changed a lot for desktop apps) and even Visual Studio 2005 needed to be patched to support Vista!
I'm currently in a 'successful company goes blind' situation myself. The company has grown massively and the situation we're stuck in is mainly driven by two types of 'internal' people:
a) People who've spent 10+ years with the company, and ended up in management/C-level positions - these are people that have:
- Been promoted over and over from entry/mid positions across a chain of smaller easier-to-deliver projects
- Have not upskilled or gained real experience on anything large/complex/challenging
- Have a very safe, very cozy job, with no perspective or understanding on anything other than their past 10+ years.
b) Technical Leads/directors who've spent ~8+ years within the company, where:
- They have a solid track record of success, a good reputation, and built up a lot of trust with the company...across a chain of smaller easier-to-deliver projects.
- From their earned track record, they have very little oversight and accountability (management doesn't think they need it)
- Limited/no interest in upskilling
- Decision making is mostly on them...and the decisions made are orientated around themselves (!)
- Limited/no interest in listening to others perspectives...even to the new highly-experienced management that's brought in to oversee them (why should they? They're the chosen! They're seen as perfect!)
You can see right between them how the blindness forms. Now, guess what happens when a client decides that a small project...is actually going to be a much, much bigger project, with real complexity, challenging external client people to work with, and a large number of external hires necessitated. Purely reactive decision making, several people that are a SPOF if they leave, no proactive planning or strategy now or before...and then things start breaking down...
The thing you have to remember with that $165k spend on tokens is that token prices are going to keep rising, and models may not get much better. I wouldn't be surprised if doing this same migration in 6 months time would end up costing $250k+
The map includes metros across the Tyne and Wear Metro in NE England, and while its not perfect, it's by far the most useful train live tracking I've ever seen. There's quite a few places in the UK with different rail systems that don't fit together (and have apps of varying quality/usefulness)
Qualcomm are slowly but steadily improving Linux support for the X1/X2 Snapdragon CPUs (such as the qcom-hamoa-ec driver in 7.2), so it's still a wait. I think there's some challenges with Secure Boot and the firmware with these Lenovo devices though.
I've been a Tidal subscriber for several years, and while I've not seen much of the AI music problem (yet), there's been a big issue with people getting their music intentionally or unintentionally labelled under other artists names. The platform has had some odd technical hiccups too over the past year, so I've started wondering how many people are left actually maintaining it all (there were layoffs pretty recently).
I used the FUTO keyboard for a few months but ditched it as the word suggestions were either odd or random (one example: I'd write 'Jordan' and it would always suggest 'Peterson' as the next word), and I'd got a weirdly passive aggressive prompt saying I should really purchase a license. Went back to the de-Googled Android keyboard in GrapheneOS
That does happen, and that junior level work is constructive, but what I've also experienced and noticed is that companies put a lot of effort into finding and hiring exceptionally skilled juniors (industrial placement schemes, graduate fast-tracks, etc.) where they can make safe bets on those people delivering significant value to their companies/projects. Some bigger companies do make an effort to build a mutually beneficial working relationship with a clear 5 year career roadmap (some sectors really struggle with those best candidates choosing FinTech instead). I've worked with high-performing junior programmers that have (quantitatively and qualitatively) dramatically outperformed experienced mid-level programmers, so I'm always an advocate for investing in them.
Your point is valid and yeah, it's a never-ending fight just to keep the control we have. Things like the Play Protect API and loads of Android apps being coupled to Play Services is it's own big challenge we're stuck with just to stay within the Android ecosystem