Well… per the linked comment’s thread… they did… by replacing the ddr chips on raspi, and raspi just did the usual walled garden corpo dystopian lockdown. For a full product with a well known name like raspi I guess it will be some years…

Any you like. I’ve had very nice results with llama.cpp w/ vulkan backend.

The rpi does not have ddr5, the price makes no sense

All the 90s kids on lemmy having a nostalgic time remembering the irc era.

Idk, I think you’re immediately jumping to the conclusions of “they’re trying to make us all rent everything”, which, while most likely isn’t wrong, is also obscuring some potential for “Hey look, I can play less-response-time-intensive xbox games on my fricking phone with good wifi, because of amazing compression technology”

How is this rage bait? It is simply properly used English language…??

Edit: To the lurking downvoters: Since when people started interpreting “mac users” to mean (according to the context here) semi-all? A qualifier of “some” i guess is what you would like, but it’s not incorrect.

It’s proper to just say “mac users”, or for example “fishermen use worms as bait”. Or maybe you’ll like this example more: “linux users write open-source firmware for wifi routers”

I am more skeptical. Even with Navier Stokes, I never heard anyone actually internalize the result and have trust in it. I just heard vague lean-dih waving.

Supposedly it’s 26000 subtheorems of goobledigook, feels untameable.

Mathematicians seemingly: Axioms are correct, we trust lean => Result correct.

As a physics, CS thinker: AI probably found a bug in our axioms. And/or usual self-referential math issue. And/or connected to Goedel’s incompleteness. Also I wonder why I even trust lean as much, havent even read its code. Most software works great until somebody starts poking at its limits.

Compared to what OpenAI is claiming its word machine has solved (without any human verification for the claims so far, not even in OAI), the scribbles kinda are (just highschool math).

I think so, yes. It bases the business on “money milking” by staggering SKUs to maximize profits. “Cheap” one that is mostly useless, just to upsell the others. “Mid” one that is just okay. And “expensive”, that isnt more expensive to make, but looks reasonable here to really milk the money. “Usual” dark patterns these days.

Per your hypothetical:

Say there is some r&d budget A, production and logistics costs B, and SKU component costs Ci. Since we’re capitalistic corpo aimed at line go up, there is also D% margins.

rough math, per unit:
((A + B) * sku_1 / N**2 + C + 2) * D = 50
((A + B) * sku_2 / N**2 + C + 4) * D = 70
((A + B) * sku_3 / N**2 + C + 10) * D = 100
sku1, sku2, sku3 = number of sku units
C - cost of all other components
N = number of units
D = 100% + margin%

Set D, A, B, C, N and you’ll see how they expect to sell their rpis, not based on “hey this is our proposal”, but based on “target this market, upsell this, and focus on selling this overpriced sku”.

But maybe they’re not evil, and they don’t want to tell you what to do, just a proposal and expected market. Let’s do a simpler situation. A per-unit equitable margin based business:

B - per unit logistics, prod costs (they are literally identical in every way except for ram chip) D < 100%, added margin

(B + C + 2 ) * (1 + A + D) = 50
(B + C + 4 ) * (1 + A + D) = 70
(B + C + 10) * (1 + A + D) = 100

Merge B + C for brevity. 1 + A + D = M

BC = 50 / M - 2
BC = 70 / M - 4
3) BC = 100 / M - 10

=> M = 10, BC = 3

Oi. Margin of 900% (that’s some costly R&D) , but 3) is literally then broken. And then math does not math, with abstract bad numbers.

Let’s try more realistic numbers:

1G - 53
2G - 87
4G - 127
8G - 194
16G - 334

BC = 53 / M - R1
BC = 87 / M - R2
BC = 127 / M - R3
BC = 194 / M - R4
BC = 334 / M - R5

Now some data is missing for ram chip prices, or costs of components, but let’s just assume indeed BC + R1 = 53/M. I.e. they don’t purposely lose money on the shittiest sku. Ram prices are exponential. Say BC + R1 = 25. Which makes M around 200%. Wait, that makes 16G chip cost at least 120. Okay, about 2.5 times as expensive than the most expensive retail-amount chip I found on alibaba.

So math aint mathing, and money appears to be a floating imaginary point in petroleum integral support space.

Oh well, companies will be companies. At least when you buy a raspberry pi you definitely are not paying them to actually specifically reduce your capabilities, gaslight you without valid engineering explanations, and just continuing to try to extinguish an unending PR nightmare using your fancy SKU cash

In reality, their base tier SKU is likely a net money loss. Which is why they’re scared shitless somebody actually broke their dark pattern. The sales department is always right on what and how many should be sold!

I swear, lemmy sublemmies should have moderators with the explicit ability and responsibility of rewriting the titles of submissions, to curate the deluge of unsubstantiated and/or egregiously misleading titles

The fact that it is economically viable to resolder different chips, is already showing it’s a illegitimate business model, imo…

Not a product, but a money grab

Have a look at Volla Plinius.

Still figuring it out, camera software definitely lacking, waydroid support still in development xD But hey, I love this thing.

While the world is getting muddled, a professional who proclaims hallucinations he doesnt know confidently is usually very quickly no longer a professional.

Exactly same thinking. I can usually type it out faster than AI (when taking into account latency, context building, prompt writing, prompt fixing, AI hallucination review) if I know what exactly needs to be done.

When you’re out of your depth, it shines in providing beautifully confident and a botanical garden of a code piece that wil most likely be broken and break in new and exciting ways. Great for incompetent sycophantically-challenged managers that forgot how to code, or never actually had any experience.

Large-scale code transformations, taking into account the bigger picture of the repository are most likely hallucination free (not generative, just transformative, as per the actual LLM model) and a very exciting use-case.

Also it’s currently a very nice pragmatic tool for checking for any mistakes, because it can connect the larger context of the repository quickly to the diff. Terrible if done by a manager without understanding of copilot and workflow, but a great tool if done through claude and cli as a pre-commit step with just quick checklists. Makes plenty of mistakes, but allows to catch your own big mistakes nicely quickly.

Somewhat tailored response per your username: Anything the ministry of truth decides should not be part of the training set.

But more generally, even in things like neurodivergence, lots of questions you could ask that would be in the bounds of “guesstimation” by professionals. The problem is these LLM will spout pure utter highly-believably-confident snake oil with the tiniest little legal microscopic fineprint on the side of the html webpage that you can never scroll to saying “AI responses can be wrong”.

For me, the mere fact that you have a nicely worded source of information that, depending if you’re lucky with your question, is pure scientific fact, or utter maddening bullshit, makes me terrified to use it.

ChatGPT and others have a ~ 90% and more hallucination rate on questions it wasnt trained on. You’re playing a game of luck using gpt and others for knowledge that you’re not verifying by source.

Always was… just had to find the seller who sells you the laptop without a prepurchased license/preinstalled windows. Sadly was always higher tier stuff that had a choice.

To your last paragraph: Yes and no. It’s not only the material that is taught, but also the approaches that you learn while dealing with it.

If you cheated on the theater exam, and masterfully didn’t get caught. On paper: great. In reality: social settings might be harder to deal with, because you would approach them in the same way. Just it’s not visible on paper.

If you didnt cheat and flunked: You’re a programmer who can stamp out code, and most likely unable to deal with anything else at work that is more theatrical than cold hard logic and math.

I always saw education (and most intelligent people do corroborate as well) as a training ground, not entirely for the substance taught, but to show that you can deal with it (your function 2, but more in the sense, given such problems, you will know how to approach them, and where to look for knowledge, because nobody remembers everything).

Ugh… How do you think they correlated the GDID to his social media?

A genz hacker. In a world where “hacking” is writing prompts and calling IT help desks.