voxelc4L
Joined 5 years ago
Karma
15
Submissions
13
Member since
May 25
Recent submissions
Comment · 3 months ago
Watching the tenants of intellectual gatekeeping bang up against the inevitable paradoxes of late stage capitalism - a spectacle worthy of jiffy pop.
Comment · 3 months ago
Hot take: Or maybe all IP should die? Maybe the moat is the effort (read: attention, energy expended etc.) per unit time to receive and/or deliver a quality product/service. Looking at IP from a perspective of pure data (or a misconstrued and narrow definition of "information") is both reductionist and ignorant.
Comment · 6 months ago
I’ve stuck to the non-1M context Opus 4.6 and it works really well for me, even with on-going context compression. I honestly couldn’t deal with the 1M context change and then the compounding token devouring nonsense of 4.7
I sincerely hope Anthropic is seeing all of this and taking note. They have their work cut out for them.
Comment · 7 months ago
Wonder how many of these cases are using the 1M context window. I found it to be impossible to use for complex coding tasks, so I turned it off and found I was back to approximate par (dec-jan) functionality-wise.
Comment · 7 months ago
The metaphor, as it relates to the Three Body Problem books, has some issues and is pretty misleading. e.g., What makes the forest dark...? It's not the act of participation, nor is it _all_ (or even a remote majority) of the participants. The issue with the forest is the imbalance. It's the lopsided distribution of ownership and power. Maybe instead of worrying about the forest we should worry about the elements of the forest that make it dark? Maybe the fact that the nations developing it are either outright communist or border-lining on some form of neo-fascism.
I don't know about you, but I like the idea of all of humanity's knowledge being constantly amalgamated and disseminated in a semantically searchable and incredibly approachable way... I just don't like the fact that it's being owned and profited off of by a minute subset of us.
Comment · 8 months ago
Not that congress or the DOJ would do a single thing about this, but here are the well defined consequences for perjury in front of congress... https://www.congress.gov/crs-product/98-807
Section 1001(a) states: "[I.] Except as otherwise provided in this section, [II.] whoever [III.] in any matter within the jurisdiction of the executive, legislative, or judicial branch of the Government of the United States, [IV.] knowingly and willfully— [i] falsifies, conceals, or covers up by any trick, scheme, or device a material fact; [ii.] makes any materially false, fictitious, or fraudulent statement or representation; or [iii.] makes or uses any false writing or document knowing the same to contain any materially false, fictitious, or fraudulent statement or entry; [V.] shall be fined under this title, imprisoned not more than 5 years or, if the offense involves international or domestic terrorism (as defined in section 2331), imprisoned not more than 8 years, or both. If the matter relates to an offense under chapter 109A [sexual abuse], 109B [sex offender registration], 110 [sexual exploitation], or 117 [transportation for illicit sexual purposes], or section 1591 [sex trafficking], then the term of imprisonment imposed under this section shall be not more than 8 years."
Comment · 8 months ago
Not sure if anyone has heard his interview on the Hard Fork podcast... was not unlike listening to a PR automaton. Now going to work for OpenAI. Yup.
Comment · 1 year ago
… or any code?
Comment · 2 years ago
It begs the question though, doesn't it...? Embeddings require a neural network or some reasonable facsimile to produce the embedding in the first place. Compression to a vector (a semantic space of some sort) still needs to happen – and that's the crux of the understanding/meaning. To just say "embeddings are cool let's use them" is ignoring the core problem of semantics/meaning/information-in-context etc. Knowing where an embedding came from is pretty damn important.
Embeddings live a very biased existence. They are the product of a network (or some algorithm) that was trained (or built) with specific data (and/or code) and assume particular biases intrinsically (network structure/algorithm) or extrinsically (e.g., data used to train a network) which they impose on the translation of data into some n-dimensional space. Any engineered solution always lives with such limitations, but with the advent of more and more sophisticated methods for the generation of them, I feel like it's becoming more about the result than the process. This strikes me as problematic on a global scale... might be fine for local problems but could be not-so-great in an ever changing world.
Comment · 3 years ago
Not that I'm arguing one way or another, but everyone posting "Hanlon's Razor, QED" should consider that Hanlon's Razor is 1) a heuristic and 2) breaks down _very_ quickly around psycho/sociopaths.
Comment · 5 years ago
Why does something have to be full-on existential threat to humanity to qualify for research/development/treatment?
Comment · 5 years ago
Is the population that visits your website more biased than average?