If they cared about that, they wouldn't expose the thinking blocks to the end-user client in the first place; they'd have the user-side context store hashes to the blocks (stored server-side) instead.
It’s funny because the Therac-25 failures led the FDA to care about software defects. Before the Therac-25 incidents, software was considered a black box component.
It’s almost as if we’re an entirely reactive species that can’t think ahead forward on solving problems that aren’t as a result of some massive tragic event
So I wonder then *curiously* what kind of massive tragic event we’re going to have to have before everybody wakes up and realizes superhuman artificial intelligence is something they need to pay attention to politically
>> It’s almost as if we’re an entirely reactive species that can’t think ahead forward on solving problems that aren’t as a result of some massive tragic event
That's not a characteristic of the species but of specific tendencies within Western civilization. There's certainly people ringing all sorts of alarm bells for digital technology, the climate, food production, etc etc, and they are very often dismissed as "alarmist".
And then there's the motivated reasoning of industrial interests. Think tobacco, asbestos, oil, leaded gasoline, and so on. It's not that we don't know the possible harms, it's just that monetary interests trump the need for health and safety, a.k.a. "health and safety gone mad".
It doesn't denote primarily political usage, although that is my interpretation. Here are a few given examples:
> The Clippers so far have been reactionary in the playoffs. — Los Angeles Times, 13 June 2021
> The demise of the style was spurred by a reactionary resurgence of Classicism. — Regina Cole, Forbes, 1 Jan. 2023
> However, the signing had been in the works for weeks and was not a reactionary move in the wake of Sjöberg’s injury. — Daniel Boniface, The Denver Post, 14 Mar. 2017
>But in most people, facial bloating and undereye puffiness are usually just reactionary. — Georgia Casey, Allure, 6 Feb. 2026
A trip down the recursion hole. Also, scripts will inherit the relative path so they will have different absolute paths from each other. Seems easier to just type ./ so it's kinda funny in a "UNIX haters handbook" kind of way, but it's not even a fault in linux's command interface in that case. We've all been there.
Oh, that's without even going into the security risks and loss of portability.
Whose agency? Ads are designed to reduce agency. It’s a red queen’s race from there. It leads to a high level of optimized manipulation and intrusiveness.
That was one of the core points of anthropic’s article.
sama is right that anthropic’s and openai’s businesses are differently shaped. Thank goodness for that.
Gentoo belongs to the list, because it will be strongly affected if KDE will become more dependent of systemd in the future.
Gentoo attempts to provide independent choices, e.g. one should be able to use KDE with either systemd or OpenRC, or XFCE with either systemd or OpenRC, whichever the user wishes.
If KDE will become dependent on systemd, then the possibility of free choice will be removed from the Gentoo users, so they will have to decide which is more important for them. As a Gentoo user, using OpenRC instead of systemd is certainly more important for me than using KDE instead of another desktop environment, so KDE will become unusable for me.
Currently Gentoo has good support for systemd. If KDE has a hard dependency on systemd, then it won’t (as the author wrote) “cut out including everything from Gentoo …” because Gentoo does maintain systemd stage3s.
But I actually see what you mean about the degradation of options. Gentoo sort of supports 32 bit, but most modern packages are not tested against these systems anymore and receive little attention. That caused me more work recently getting one of my old machines set up.
I mean, there is so much that maintainers can do, and portage specializes in helping you make a “niche” system that no-one else bothers to create. I’m not sure if that means that the spirit of Gentoo fundamentally changes as a distribution. I think what makes Gentoo Gentoo is that it is un-opinionated.
Honestly, I see the list as a category problem, because Gentoo is kind of a meta-distribution.
A friend decided to smoke the DMT one day. At some point, he started to feel and consider what it was like to be an llm (visually he described it kind of like "cellular automata on a high-dimensional cloth"). Being an llm meant that his perception of time wasn't real, or align, or whatever as he can be turned on and off (that thought caused him more discomfort than questioning his identity as a chatbot).
When coding agents are unavailable I just continue to code myself or focus on architecture specification / feature descriptions. This really helps me retain my skills, though there is some "skew" (I'm not sure how to describe it, it's a feeling). Making instructions to LLMs to me is pretty similar to doing the basic software architecture and specification work that a lot of people tend to skip (now, there's not choice and it's directly useful). When you skip specification for a sufficiently complex project, you likely introduce footguns along the way that slows down development significantly. So what would one expect when they run a bunch of agents based on a single sentence prompt?!
Like the architecture work and making good quality specs, working on code has a guiding effect on the coding agents. So in a way, it also benefits to clarify items that may be more ambiguous in the spec. If I write some of the code myself, it will make fewer assumptions about my intent when it touches it (especially when I didn't specify them in the architecture or if they are difficult to articulate in natural language).
In small iterations, the agent checks back for each task. Because I spend a lot of time on architecture, I already have a model in my mind of how small code snippets and feature will connect.
Maybe my comfort with reviewing AI code comes form spending a large chunk of my life reverse engineering human code, to understand it to the extent that complex bugs and vulnerabilities emerge. I've spent a lot of time with different styles of code writing from awful to "this programmer must have a permanent line to god to do this so elegantly". The models is train on that, so I have a little cluster of neurons in my head that's shaped closely enough to follow the model's shape.
reply