snvzz

And, of course, the current solution is formally proven fully reliable and has all these certifications.

(in which case, why replace)

Are we talking about the same thing?[0]

0. https://github.com/keirf/flashfloppy

If the intent is to play with RISC-V cheaply rather than getting the absolute cheapest computer, I'd look at something with C908 (a small cpu that is vector capable, RVV 1.0).

A particularly good deal is the Banana Pi BPI-RV2: It has 4xC908@1.25GHz, 512MB DDR3, M.2, miniPCIe, 1x 2.5GbE, 5xGbE.

There is a cheaper Banana Pi BPI-CanMV-K230D-Zero. 2xC908, 128MB LPDDR4, KPU, MIPI CSI/DSI, Wi-Fi, 40x GPIO.

If serious about RISC-V, then of course RVA23 is desirable. Unfortunately any SBC with that is currently expensive due to the RAM shortage situation.

Replacing the drives with FlashFloppy would not be a crazy idea as a first step.

That's how you get the law enforcement knock on your door.

Good old iCE40 and ECP5, then.

Not many options unfortunately. Today that's GateMate.

Next best thing? The mature reverse-engineered families nextpnr[0] supports.

0. https://github.com/YosysHQ/nextpnrhttps://github.com/YosysHQ...

And so many more are hidden in there, in those MLoCs.

The only way forward is a better system architecture, as enabled by the seL4[0] microkernel.

Any solution involving keeping Linux or the monolithic kernel is just stopgap or outright delusional.

0. https://sel4.systems

Likely shipping w/o the proper documentation to enable nextpnr and similar tools, as it's tradition in altera.

I will always choose the vendors that do things properly, when they're even an option.

Neat, except power supply (a very important part) is being ignored.

The way forward is to remove contracts and try again.

The proponents of Contracts can then polish their proposal and try again.

There's value in protecting stability and the status quo, dodging the mistakes that countries which quickly embrace things make.

RISC-V is inevitable.

I like 0BSD, except the name.

The proliferation of BSD licenses dilutes the meaning of "bsd license". There's only one MIT license, so there's no confusion.

Too bad it's not open source.

("NC" creative commons)

AIUI there's no "public domain license". Permissive licenses like MIT are as close as it gets.

And the disclaimer of liability is kind of important.

The mandatory Bone Song[0].

0. https://www.youtube.com/watch?v=kHgRlsOvN3M

>Copyleft.

Why even bother with the extra "protection". It can be re-made by AI as easily.

Just go all the way, MIT or BSD. Maximize reusability and adoption.

Now, for contrast, I'd love to see what Opus 5.5 and/or Astra can do when set to do the same thing on a 52m wall clock limit.

Tailscale shows up.

Whenever I need tickets it has been a smooth experience via ekinet[0].

Just entering a code later in a ticket machine was enough to get it printed.

0. https://www.jreast.co.jp/multi/en/

>How a $500k/Year AI Engineer Lost Everything

Apparently they saved no money and went homeless in a matter of months (?!).

People who work in mcdonalds putting the fries in the bag do not earn 500k/yr and somehow manage to avoid homelessness.

It is quite telling that they blame the AI rather than themselves.

Crazy idea: Putting them in Prison, so that the rest of us can live in peace.

There's no end to bugs like these in Linux. Millions of Lines of Code can easily hide many, many of these. Code that runs in supervisor mode.

None of this can ever affect seL4[0] nor a system built on it, due to its architecture.

0. https://sel4.systems/

Found the actual hypervisor[0]. It is GPLv2.

0. https://github.com/losfair/l4h

It would have prevented it. e.g.

>Of course, you’d want to try ../… Yes, it does let you escape the temporary folder...

This is absolute bull on seL4 by construction. "sandboxes" that prevent access that the OS otherwise grants. In seL4, this component would not need a sandbox in the first place, because it would never have a capability to some VFS path it doesn't need to access.

This sort of escape relies on ambient security, which is the security model of UNIX and Windows NT. Access is granted by default and then restricted, rather than simply not granted. A broken security model.

And, unlike a demonstrated shitty browser sandbox, seL4 formally guarantees isolation. Capabilities cannot be forged.

Who voted Claude in?

seL4 can guarantee* time domain isolation between processes.

*on some CPUs. There's a WIP time fence specification in RISC-V to try and improve this.