That's right: no legal basis exists now. The proposal is to create such legal basis.
And if "No large undertaking could ever function with such broad exposure to liability" - that would be great, i think we would prefer that such firms doesn't exists.
> that would be great, i think we would prefer that such firms doesn't exists
They stop existing within your jurisdiction. Also, the idea that the public would go along with any of this for this issue is silly. Let's start with crimes that actually cost lives.
I'm hosting few hobbyist public websites. One is the library of hymns, songs, ... - mostly PDF, MP3, MIDI.
OpnSense router -> Traefik -> Docker container.
That Docker container runs on NAS, which is connected via 1Gbps port.
Suddenly, last week I started to get over 1000 RPS. Was it scanner, parser, crawler, DDOS? I can't tell, as user agent was default Mac OS.
It was using ~1Gpbs bandwidth. From many IPs.
If I would have it hosted at AWS / Azure, it would cost me > $40 / hour for the bandwidth. Unless I use some of their CDN.
My router CPU went high, but it handled it ok.
NAS CPU went high, but it handled it ok.
Traefik loaded the server heavily: it had to both receive and send 1Gbps. And write access logs for all requests. My Alloy & Loki logging system got loaded heavily.
correct me if I'm wrong but wouldn't a Thunderbolt 5 cable $79 from Apple[1] work better for that? It gets 80 Gbps and doesn't need any extra equipment.
We refurbish laptops and hence distribute .iso / ipsw images to various refurbishing stations. (or also, think of PXE/network boot situations). More images x more laptops refurbished at once = more bandwidth.
Obviously it's possible to move this to a NAS, not use a mac, etc. etc. - but just wanted to chime in with my little use case of moving multi-gig-files around. :-)
I was just looking for a thunderbolt-sfp solution. Super timely!
> Author talks about 4k video, but AI tells me that for "4K production mastering (ProRes 422 HQ, 4:2:2, 10-bit)" the rate is under 2Gbps.
That'd be the rate for 1 stream at 1x playback. Multiple streams because of overlays/crossfades/etc... along with playback at >1x (such as during rendering in particular) will change that significantly.
NAS stuff can benefit greatly and straightforwardly.
A typical NVMe SSD can easily do sequential reads at 7GB/s, which outpaces a 25Gbps connection by more than 2x. NAS is more typically backed by SATA HDDs, but there's some relatively easy way to put an NVMe disk in front of them as a cache.
Do you ever _need_ it? Perhaps not, but there's plenty of common activities just stemming from NAS usage which see a linear speedup, and would continue to get faster up to 100Gbps even.
ProRes is a compressed and encoded format. In order to become ProRes, it needs to be uncompressed and decoded at some point. uncompressed and unencoded is easier to work with on the rendering side, but much heavier on the data side. Ask your AI about the DPX file format.
Normal 4K editing workflows are fine with even 2.5 Gbps wired networking (wired to make latency like when scrubbing not-horrid)... 10 Gbps more necessary if editing ProRes RAW and doing multicam (4-8 streams simultaneously).
I'm in a never-ending pursuit of cutting time when transferring large files though — AI models is one thing, but I have 100-200 GB projects I transfer across for backup or import (using an external TB5 drive capable of 3-5 GB/sec), and if I can do that in 1 or 2 minutes per project instead of 3 or 4, that's a noticeable speedup.
Definitely overkill for 98% of my workflow though. 10 Gbps is fine.
I’ve been running a Supermicro X10SLH-LN6TF as my router, and it’s very well suited for the job. Power use is very high for a home router, but reasonable for a 1U router - especially considering the era it’s from and its capabilities (routing 5-8Gbps per stream depending on firewall config, able to handle a couple active streams, and line rate routing with no firewall).
I wouldn’t suggest anyone buy something like this today for home use, but 6 years ago when I purchased it there were not many 10GbE routers to be had period, let alone for the price ($300 roughly if I remember correctly ). Even considering the cost of power over the last 6 years, I still come out much cheaper than if I’d bought something more power efficient new at the time.
FWIW, I got 10 Gbit at home 2 years ago and built my router for $100 using a ThinkCentre M720q Tiny running on a Core i3-8100T (a mini-PC from 2019).
It draws around 15W idle, going up to 30W at full bore. For comparison, the 10G rental router from the ISP drew 15W no matter what (idle or full bore routing). This is with a 10GBase-T SFP+ which probably influences the power draw vs using a DAC.
It has no issues maxing out at 9.4 Gbit routing with a single or multiple streams of iperf3 with your basic IPv4 NAT+some port forwards config.
It should be doable by most software engineers. They did teach some basic CCNA course in the university as part of computer science studies and the concepts taught there should help a lot with configuring any router, including RouterOS. It could also be helpful to watch some guides on routing and firewall on Linux, since RouterOS is just Linux with a very specific UI on top.
There is a safe mode built in. If you enable it and you loose connection to the router due to some misconfiguration, it will reset within 15 minutes to the point of where you started the safe mode session. Also if you mess up the IP layer, there is a chance that you can still connect and configure RouterOS through their WinBox program using the mac address.
Yeah, it's inconvenient when the [effectively] black-box router device that only the bot understands goes out to lunch, and takes access to the bot along with it.
But the degree of inconvenience is up to the beholder.
One way is using something like ChatGPT with a pocket supercomputer. It's tedious to get technical stuff done that way, but it can be accomplished. (For those who haven't tried it, the bot can be spooky-good at correctly interpreting cell phone photos of technical text on a screen. This saves a ton of transcription effort.)
Another way is with the phone's hotspot and just about any random computer. Laptops have wifi. Wifi dongles for desktops can be stupid-cheap, and it's useful to have a few scattered around. It works fine (unless one's hotspot data is ridiculously-limited, I guess).
The best way is probably by being prepared with procedural fall-back methods. Keep backups of your configs, and keep the recovery instructions with them. Test recovery; an untested backup isn't necessarily better than having no backup at all. Keep as many historical versions as one wishes, and automate it because procrastination is part of the human condition.
After that: When the router's config gets stuffed up, then the backups and the restore instructions are right there. Just pick an appropriate backup and use it.
We should probably be keeping backups of important stuff like that even in the absence of LLMs. Humans also err, and stuff does break, and what seemed obvious six months ago when we were in the groove might be impossible to remember when we need it today.
And if "No large undertaking could ever function with such broad exposure to liability" - that would be great, i think we would prefer that such firms doesn't exists.