They haven’t activated Astra for me yet, I have two resets now. I’ve been using the opportunity to test out how good 5.6 Sol is at computer use asking it to generate stuff in Blender which has been… interesting
My issue with OVH is the support is God awful and half French. Someone made a complaint about my server and their default stance was to nuke me from orbit and ask questions later.
Oh and their "data centers" are just shipping containers that are prone to the occasional fire.
Much happier with Hetzner, even though it's more expensive.
The RIAA still has insane amounts of power both political and actual. They got away with successfully suing an ISP for billions over of what their customers downloaded and only lost on appeal. They've shaken down several other ISPs for settlement money to avoid going to court. Lately they've also been going after stream rippers, AI start ups, and archive.org
I don't that's a fair description of either 'next-token predictor' or 'stochastic parrot'. Both of those terms describe mechanism, not value--the fact that people squawk that the terms are minimising is projection on their part, not inherent to the phrase.
What is intrinsic difficulty? A product at scale has many intrinsic dimensions, not just technical. Saying scaling is not an intrinsic difficulty is pretty weird given a highly scalable product usually looks nothing like their 1-user counterpart even when they have the same functionality.
Actually, the moat is regulatory. Expect these companies to behave themselves in progressively more grotesque and sycophantic ways to get the federal government to make open/foreign models (and their output) illegal.
It looks more like awards for events, e.g. "Most Creative and Engaging Event" and "Best Event Series (over 1800 attendees)," none of which seem to have anything to do with the quality of any awards given at those events. What am I missing here?
Anything where you use a state machine that has to remember the parent state is effectively an intrusive linked list (even if they rarely have more than two elements); similarly the most obvious implementation of undo/redo. In these uses it's less complex and more understandable than having an explicit container, and constant-factor performance is irrelevant since we're talking about spending a few clock cycles to retrieve information in response to a human-speed GUI interaction.
You might also note that TFA is much more recent than "the end of the 1990s" and describes one of the most important software systems out there, which to the best of my knowledge still uses these techniques in the same way.
>> Along the way, it wrote 13 million lines of Lean and proved 29,500 intermediate theorems.
> Pretty insane.
I don't think the count of "intermediate theorems" tells you anything. Here's something from an algebra textbook:
---
Let G be a group, let H be a subgroup [of G], and let N be a normal subgroup [of G]. Then
H ∨ N = HN = { hn | h ∈ H, n ∈ N }.
---
This says that the subgroup closure of H and N, the smallest subgroup that contains them both, is identical with the set consisting of all products of an element of H (on the left) and an element of N (on the right).
Part of the proof:
---
Suppose that x and y are elements of [the set of products hn]. Then x = h₁n₁ and y = h₂n₂, where hᵢ ∈ H and nᵢ ∈ N. Now h₂⁻¹n₁h₂ = n₃ ∈ N, as N is normal in G. So n₁h₂ = h₂n₃. In this case
This will translate directly into lean. If you do it this way, you will prove at least 10 of what would be described in lean as 'intermediate theorems':
∃ h₁ ∈ H, ∃ n₁ ∈ N, x = h₁ * n₁
∃ h₂ ∈ H, ∃ n₂ ∈ N, y = h₂ * n₂
h₂⁻¹ * n₁ * h₂ ∈ N
n₁h₂ = h₂n₃
xy = (h₁n₁)(h₂n₂)
(h₁n₁)(h₂n₂) = (h₁(n₁h₂)n₂)
(h₁(n₁h₂)n₂) = (h₁(h₂n₃)n₂)
(h₁(h₂n₃)n₂) = (h₁h₂)(n₃n₂)
h₁h₂ ∈ H
n₃n₂ ∈ N
But none of these would be called an "intermediate theorem" in a paper proof.
Infinity project (game) https://gamedev.net/blogs/blog/73-journal-of-ysaneya?page=12 started in 2004. The developer still seems to be going for something related, it's not the same game. I don't remember how it's called, but I think it's on Steam as alpha release. That journal is very interesting btw.
There's a situation like with Amdahl's law here, which I'm guessing you're familiar with since this is HN but if not then I'm sure you can read about it. Consumer prices - the thing you're experiencing - have multiple inputs, and wholesale price is the only one that's subject to this change by generation technology so that dilutes the consequence of any such change for consumers.
If you're paying 25 cents per kWh and and the wholesale price was $100 per MWh then 10 of those 25 cents were for wholesale electricity, but you're paying 15 cents per kWh for "other things". Maybe a huge improvement to generation technology reduces that wholesale price to $90 per MWh, amazing, but alas general inflation increases your suppliers other costs to 16 cents per kWh and so you pay... 25 cents still.
If that wholesale price somehow halved and it was all passed on (good luck with that) your 25 cents per kWh goes to 20 cents, only a 20% drop.
The learning curve for solar is shallowing, and while it's steeper for offshore wind and especially floating offshore wind, those are both way more expensive than solar, so you'd need much more price decrease for them to be cheaper than say, coal.
> Why are you giving the AI a harness and a plethora of tools and not the human in this comparison?
I am not. Multimodal models generate that output directly, without tools. Which 'tools' does AI use to generate all those images, songs, and videos do you think?
> Also, that situation is extremely contrived. If a criminal threatened to kill me if I misspelled a word, I would choose a dictionary.
Of course it is contrived, it is a thought experiment. Does not make it less valid. It is essential that you don't know what task it is going to be, just that it is a task requiring a lot of intelligence. This way question dodging loopholes like "I'd choose a dictionary" are impossible (and people will always try to find some cheesy exit rather than facing reality). You have to commit to something or somebody that has broad and general intelligence; you do not have the luxury of choosing the perfect tool for a very narrow task.
> For all real precarious dangerous situations, I would obviously choose a human.
OK, so the task ends up being to write a 40-page analysis of the result of a specific experiment in quantum chromodynamics, to be finished within 2 minutes. Does your choice for a human work, or would you have rather chosen any of the frontier AI assistants? Be very, very honest. Remember that the lives of your family are on the line and nobody will judge you for a lack of allegiance to humans.
Again, don't go for shitty loopholes. Engage with the thought experiment in good faith and thus as it is stated, not some conveniently distorted version of it.
> and those meatflaps are normally called vocal folds/cords btw
What? Next you're going to tell me that meatspinner is also not the name for the human male reproductive organ.. Maybe I need to get a refund on my Temu Gray's Anatomy.
The LLM does not determine the next token. It generate odds for all of the tokens it knows as to their likelihood of being 'next'. It's up to the harness running the LLM (and in most cases the a temperature setting) to actually decide on a particular next token. I think it's more accurate to call the thing the LLM actually generates (an ensemble of probabilities) a 'prediction'. It might be accurate to say the harness decides on the next token based on the prediction from the LLM. The role of the LLM is much more akin to predicting your opponents move than deciding your own.
