Not talking about VRAM or unified memory. Regular DRAM.

There's been a lot of advancement recently on how to best manage MoE expert swapping (FreeToken is just one example).

How are you connecting to Slashdot from the year 2024?

If you have ~128 GB of RAM, consider Qwen 3.8 Flash + Freetoken. You'll only get half the tok/s as 27B but the quality is significantly higher than 27B - it beats it solidly in all compared benchmarks (for example, DeepSWE 1.1: 42,2% vs 58.7%).

Grok is a "discount model". It's cheaper than Claude Sonnet yet better than it. Doesn't compare to Fable or Opus, but then again, they're way more expensive models.

Note: I avoid using Grok - not just for the Musk factor, but because there are better options, e.g. GLM-5.3 and Qwen-3.8. Also, if you really want to save money, GLM-5.3 Flash and Qwen-3.8 Flash are amazingly good for their price range.

As someone who just had FTTP (yep, to the premises) installed in their rural home, the way they did it was thus:

- They went to the local cabinet (which is already fibre-backed to the local exchange many km's away), strung a fibre out of it.
- They got up the telegraph pole in my streets, strung a new fibre off it.
- They used the old copper line that's been there since the 60's to slide the new fibre reel down.
- They put a second hook on my eaves, and hooked the fibre from there, and then wired the fibre direct into my house.
- They didn't remove a damn thing.

Took about 20 minutes, judging by the CCTV.

It would take about the same time just to pull down the copper from my house safely (the above did not require closing or working in the road, for instance), then isolating that from the dozens of others in the road on the pole, then removing that all the way back to the cabinet. Which I know goes underground because when I first moved in a few years ago, the BT guy had to remove that cable from, basically, a huge puddle which was cutting my DSL speed from 75Mbps to 20Mbps (DSL is amazing, though, that it still worked!).

Until everything's fibre, you can't remove those old lines, it's just not practical. And then are you really going to mobilise a bunch of people to remove a small run of copper from every street?

I once did some maths and the IT at the school I work in has something like tens of thousands of km+ of copper wiring that's unused. One room can have 30+ runs of nearly-100m. That's 3km for one room, from where a building was previously wired to have an IT suite, but then changed purpose, etc. We could, in theory, take that pure-copper cable out and recycle it. That one school likely has more copper in it than my entire village.

The fact is that it's just not worth the effort. There are huge thick bundles of it, and while it wouldn't take that long to pull it all out (much easier than putting it in because you can largely just cut and pull the entire bundle), it's not worth doing so in terms of the cost of the raw copper. We know that because honestly much of the time, rather than re-use the old copper runs which we know to be Cat5e, we just run a new Cat6a run alongside them if we do need to wire something. And Cat5e has far more metal in it than old telephone line.

I cut something like 200 cables this summer alone. And I mean I cut the last 2 metres of them because they were in my way in the cabinets. The actual RUN of cable is still in the walls and going all over the building, 50, 60, 70m runs all over the place.

I bet that a single office or school demolition recovers more copper (and steel and other stuff) than an entire village of old telephone lines, and while it may be a factor on the demolition companies finances, it's not a huge one. Scrap metal, which you have to de-sheath, transport, etc. isn't as valuable as people make out. It's why nobody arranges a regular scrap-metal collection at my local council to come and just grab anything of value that they could. No. They do that for paper. They don't do it for metal unless you arrange it with them. There's not much money in it. The only people making money are the guys lobbing some bits of lead, steel and alumium on the back of their van until they have enough to get a couple-of-hundred-quid free personal backhander as part of their work when it's all been moved at the expense of their company's fuel expense account.

Yeah, all you need is 200 years for them to book enough appointments to actually pull the stuff down, and you're golden! (or at least, copper-coloured).

Like an online bookseller turning into a global retail warehouse then renting out their datacenter services, producing talking home assistants, and launching rockets into space?

[ConfusedMathLady.gif]

Ed Zitron is a moron who has predicted the past nine out of zero AI crashes, who simultaneously gets paid to hate on AI while also getting paid to promote AI (DoNotPay).

The funny thing is that he openly acknowledges that inference is very profitable (~40% margins) for all parties - he just argues that you should count training. But that argument makes no sense, because if capital tightens across the AI industry, then everyone will cut back on training and scaleup, and... then what? Where's the implosion? Everyone is profitable then. You may lose certain players in competition against others (their as-mentioned profitable assets being gobbled up by others), but the industry remains healthy.

There are more dramatic types of possible contagion risk, but they all end up with *someone* holding onto the high-profit-generating inference assets, even if it's high-seniority investors, leaving the the low seniority investoors high and dry.

The enshittification of Cursor continues.

Maybe this will force them to support more open models.

Meanwhile, what those evil liberals think should happen to YOU is... that you be given healthcare and a general social safety net.

They've been launching dummysats from Starship for a number of launches now, so I'm not sure why you think they're "forever away". The only reasons they've been using dummysats and not actual satellites is 1) Not wanting to risk more valuable payloads thusfar, and 2) wanting to keep Starship suborbital (e.g. trajectory that reintersects the atmosphere) at just under orbital velocity until they can verify deorbit reliability (failure to do so would leave that massive ship falling onto a random point on Earth).

If they were to choose to, they could start launching payloads on all subsequent Starship missions, even if they're not yet ready for reliable recovery or quick reuse. In fact, given the killing off of Falcon 9 in Florida, one can presume that's the plan (if not immediately, then soon).

Um... no they're not?

Further reading: Musk's other unstated reasons for shutting down Falcon 9 in Florida (Behind The Black)

TL/DR, the author thinks that SpaceX shutting down Falcon 9 launches from Florida was secretly a magnanimous gesture by a charitable Elon Musk to boost the competition:

I therefore honestly believe that SpaceX’s decision was not entirely based on its own internal needs. Those needs certainly were the main motive, but based on Elon Musk’s passionate dream of making the human race multi-planetary, as well as his repeated willingness to help his competitors when they needed SpaceX to launch their satellites, I believe he recognized that SpaceX’s dominance could be hurting the rest of the rocket industry, and that the instant it made business sense to clear the field for these competitors, he was quite willing to do so.

LOL.

Yes, I remember way back when Mike Pence had nearly absolute unchecked authoritarian power.