Star Trek ODN technology finally here?

[AVGresponding]

https://www.akhetonics.com/

[Haenk]

All in all this seems like a serious project. Not belonging to the "dodgy" category, IMHO.



https://www.xing.com/profile/Michael_Kissner6
https://www.researchgate.net/profile/Leonardo-Del-Bino

https://pitchbook.com/profiles/company/490660-30

https://www.sprind.org/de/projekte/akhetonics/

https://intelignite.com/companies/akhetonics/

[madires]

An optical NPU (network processing unit) and DSP world be quite interesting for telecommunications.

[AVGresponding]

All in all this seems like a serious project. Not belonging to the "dodgy" category, IMHO.



https://www.xing.com/profile/Michael_Kissner6
https://www.researchgate.net/profile/Leonardo-Del-Bino

https://pitchbook.com/profiles/company/490660-30

https://www.sprind.org/de/projekte/akhetonics/

https://intelignite.com/companies/akhetonics/

Agreed, but some of their claims seem somewhat... optimistic.

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

[tggzzz]

All in all this seems like a serious project. Not belonging to the "dodgy" category, IMHO.



https://www.xing.com/profile/Michael_Kissner6
https://www.researchgate.net/profile/Leonardo-Del-Bino

https://pitchbook.com/profiles/company/490660-30

https://www.sprind.org/de/projekte/akhetonics/

https://intelignite.com/companies/akhetonics/

Agreed, but some of their claims seem somewhat... optimistic.

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

Ah, that's more useful.

I wonder why they limited themselves to the remnents of 1960s that is still around, i.e. bitslice and TTL. I would have thought that bit serial implementations would suit optical technology.

I know of two bit serisl.computers. The Elliott 803B, as can be seen running at TNMoC, and an obscure Decca Navigator airborne tactical navigation computer the size of a couple of shoeboxes.

[Haenk]

Agreed, but some of their claims seem somewhat... optimistic.

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

Being pessimistic when trying to arrange funding money is probably not helpful
After all they do have some working silicon, so that is something to be proud of.

[AVGresponding]

All in all this seems like a serious project. Not belonging to the "dodgy" category, IMHO.



https://www.xing.com/profile/Michael_Kissner6
https://www.researchgate.net/profile/Leonardo-Del-Bino

https://pitchbook.com/profiles/company/490660-30

https://www.sprind.org/de/projekte/akhetonics/

https://intelignite.com/companies/akhetonics/

Agreed, but some of their claims seem somewhat... optimistic.

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

Ah, that's more useful.

I wonder why they limited themselves to the remnents of 1960s that is still around, i.e. bitslice and TTL. I would have thought that bit serial implementations would suit optical technology.

I know of two bit serisl.computers. The Elliott 803B, as can be seen running at TNMoC, and an obscure Decca Navigator airborne tactical navigation computer the size of a couple of shoeboxes.

Probably so as to not try running before they can walk. I think the paralleling is more optimal than serial, given the issues with high speed clocking for sync purposes, and the fact that optical leads itself very nicely to multiplexing using a single channel. The reversible computing idea is quite interesting.

Being pessimistic when trying to arrange funding money is probably not helpful
After all they do have some working silicon, so that is something to be proud of.

Possibly so, but there's also the risk of being accused of underdelivering and losing your funding stream...

I do wonder how scalable their working silicon is.

[tggzzz]

All in all this seems like a serious project. Not belonging to the "dodgy" category, IMHO.



https://www.xing.com/profile/Michael_Kissner6
https://www.researchgate.net/profile/Leonardo-Del-Bino

https://pitchbook.com/profiles/company/490660-30

https://www.sprind.org/de/projekte/akhetonics/

https://intelignite.com/companies/akhetonics/

Agreed, but some of their claims seem somewhat... optimistic.

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

Ah, that's more useful.

I wonder why they limited themselves to the remnents of 1960s that is still around, i.e. bitslice and TTL. I would have thought that bit serial implementations would suit optical technology.

I know of two bit serisl.computers. The Elliott 803B, as can be seen running at TNMoC, and an obscure Decca Navigator airborne tactical navigation computer the size of a couple of shoeboxes.

Probably so as to not try running before they can walk. I think the paralleling is more optimal than serial, given the issues with high speed clocking for sync purposes, and the fact that optical leads itself very nicely to multiplexing using a single channel. The reversible computing idea is quite interesting.

Indeed.

Serial computers were used because they use fewer active components, at the cost of speed. If the interesting and difficult bits are optical and fast, I wonder why serial wasn't considered. It could simply be ignorance.

[JohnG]

Possibly so, but there's also the risk of being accused of underdelivering and losing your funding stream...

The key here is to greatly promise and get a sh*t-ton of funding because you will change the world with your new tech. Then when you fail, fail big! Blame your funders for not sticking with the project! They don't have the vision and are not worthy of the shiny future!!!

Remember Putt's Ploy: "If you must fail, fail big."[1]

EDIT: I am being harsh about the marketing. The work itself is interesting.

John

[1] Archibald Putt, Putt's Law and the Successful Technocrat

[coppice]

I know of two bit serisl.computers. The Elliott 803B, as can be seen running at TNMoC, and an obscure Decca Navigator airborne tactical navigation computer the size of a couple of shoeboxes.

There have been quite a few bit serial machines, many you may not be aware were bit serial. For example, the PDP8 came in various forms, the low end of which was bit serial. Then there were various massively parallel machines, with each node being bit serial, like the MPP or the Aspro. Something like the PDP8 was limited in flexibility by the need for compatibility with parallel implementations. The MPP and Aspro could perform variable length arithmetic in one bit increments.

[coppice]

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

When I read that file with Chrome on Fedora all is well. When I read it with Firefox on Fedora the pages keep reordering as a scroll up and down the file. Is anyone else seeing that? I'm wondering if its a quirk in the file, or a bug in Firefox. Firefox reliability has been going downhill badly in recent months.

[AVGresponding]

Incidentally, here's the paper behind the marketing wank: https://arxiv.org/pdf/2403.00045

When I read that file with Chrome on Fedora all is well. When I read it with Firefox on Fedora the pages keep reordering as a scroll up and down the file. Is anyone else seeing that? I'm wondering if its a quirk in the file, or a bug in Firefox. Firefox reliability has been going downhill badly in recent months.

  I read it on Firefox just using the browser pdf function, without any issues