_petronius

> A conviction that the future is bright isn't enough to actually make the future bright.

Well said.

We have to be curious, open, and ready to use the benefits of technological change, but also skeptical, concerned, and willing to defend social and political norms, too. Wild optimism is potentially very dangerous; hardline pessimism would leave us insufficiently ready to face the moment when we do need to act proactively.

The other EU-level regulation that needs to be either removed or completely rethought (since it will clearly not be enforced in a way that makes sense) is the cookie regulation. It was well-intentioned, badly implemented, and the GDPR addresses more of the core problems, it is time to do away with it.

But as a whole, I push back against the idea that deregulation is the primary way in which the EU can or should become competitive with the US on technology. Lack of public investment, worse ability for companies to offer equity incentives, and timid private investment are all much bigger problems than consumer protection regulations.

It's a pretty hopeful, optimistic view of the future in a time of high uncertainty, but also represents a positive argument: it's worth doing these things because they are interesting, weird, and fun, and because they represent a continuity with past and future people we will never meet.

Plus, you can already buy a ticket to the finale, so your distant descendants can go see it :)

1. Free speech in the public sphere includes the right of people offended by your speech to express that they are offended (to speak back). People like Toby Young (founder of the FSU), and those who decry "cancel culture" do not like this, but it is part of what constitutes genuine free expression.

2. Free speech against the government, which the government should not be allowed to punish people for, to enable a pluralistic and democratic society. E.g. someone like Mahmoud Khalil who was recently arrested for his political views on Palestine (in violation of both the principles of free speech and his rights as a green card holder).

People angry about #1 (arguing) tend to conflate it with #2 (genuine censorship), and accuse people who accuse those approximately equal (or less) in power than them as equivalent to draconian speech-controllers. People angry about #2 are pointing out a genuine power imbalance, and the damage to democratic society.

Elon Musk embodies this perfectly: he argues in favor of free speech when people tell him his opinions are objectionable, but actively censors speech on Twitter that he doesn't like. In one situation, he simply wants people to receive his speech without speaking back, in another, he wants to stifle discussion.

Speech is not a unidirectional activity. In order for it to be valuable and protected, the act of speaking in public is the act of listening to criticisms of what you say.

(Also, Trump did not run on these policies -- he's implementing policies he explicitly repudiated when he lied about not being affiliated with the Project 2025 manifesto.)

I don't want to let the voters off the hook, but continuous, ongoing responsibility for this unfolding economic and political disaster lies with the people who hold institutional authority, and are refusing to use it to comply with the law as it stands.

No electoral mandate (and the argument for a clear mandate for all of this is thin or nonexistent) makes unconstitutional/illegal action suddenly legal or constitutional.

Whether anyone with the relevant power chooses to punish these violations, is a different matter. The choice since January 2020 has been to repeatedly do nothing in the face of illegal action, but winning elections doesn't make criminal action magically non-criminal.

The gold standard, IMO is something like the TNG episodes "The Best of Both Worlds" pt 1 and 2 -- an end-of-season cliffhanger that rewards you returning to the show by telling you what happens next!

I think the lacuna here is meant to add to the tension and mystery, but I agree that the new season has started off frustratingly slow. You gotta wrap up stuff to move forward with a plot, otherwise it's all just treading water for the sake of atmosphere.

An acknowledged, well-designed, and state-supported path to vocational education is very good; social mobility is important within such a system, and a lack of social mobility doesn't have to be baked in.

Walking through mockups and saying "ok, but what happens if I do this" is essentialy the same question as "what are the acceptance tests for this?"

Often that happens in various kinds of meetings over the course of feature development, when it would be more effective to just immediately create the tests up front, which can then go and be developed against. With current tooling, it's actually possible to get the PMs to participate in detail in writing tests. Although that point is often where I have gotten the most pushback.

Makes me wonder how hard it would be to show things like historical coastlines in England and the Netherlands, or historical watercourses, but I guess that could be both hard to visualize, and you'd have to compile that data from a lot of different sources.

I have two nitpicks with this type of view of historical world maps (not this project specifically, it just employs a visual vernacular that I have opinions about):

1.Drawing a border around an area and shading it in doesn't mean the same thing in all times and places. It might be a state with a central government as we think of them now, or it might be a collection of states or proto-states that are conventionally grouped by common features of their cultures, or it might just be an area where the pottery is consistently similar.

2. More importantly I think the areas _outside_ the shading can be misleading, too: it makes the world look empty, even though most of the world (but not all! especially in the places settled by the Polynesians much later) definitely had people in it by the time this timeline starts.

It gets a little far into the weeds in a very "the map is not the territory" kind of way, but is fun none the less.

I think (it is not addressed to my satisfaction in the video) that it is implied that from inside the black hole, assuming it was formed from a collapsing star, you would see all around you the event horizon: behind you, the unverise at the time you crossed the event horizin, in front of you, the surface of the star at the moment of its collapse. The singularity, as the post covers, exists only in your future, so you could not see it, even though you will always end up there.

Certainly there is a trend towards this in a lot of industries besides computers, but given how powerful and expensive these devices are now, the current upgrade cycles are crazy fast. I think consumers are souring on them a bit as well (both because of the price, and because the annual new models have really slowed down in the visible feature improvements they offer).

I know the economic incentives for the producers are aligned towards repeated purchases, and that's super tough to realign, but how long can the market and the environment support four-digit phone price tags that are upgraded every 1-3 years?

First, there are repeated documented examples of prompts beign designed that can cause an LLM to output training data (first link from a quick google: https://www.darkreading.com/cyber-risk/researchers-simple-te... but there are other examples, and some well-discussed ones on this site involving Github's own Copilot).

Second, the "it's just like a person learning" argument has been applied to all sorts of machine learning, and it rests on several fallacies:

1. That these systems learn the way humans do, and innovate on that learning

2. That their output constitutes any kind of original thought (related, LLM output is not copyrightable; human output is)

3. Most importantly, the scale is totally different. I think we can agree on the trivial example that training an image generation LLM on an artist's style and using that at scale to undercut the market for their work would constitute a kind of technologically-enabled competition that normal humans learning and copying styles could not equal, either in speed or in cost -- even if the quality were orders of magnitude better.

I actually think there is an attempt in the ToS as written to keep the definition narrow in the way that the parent comment quotes, but it is unfortunately not super well drafted, because there is a partially-explicit distinction between "platform" and "services" that may make sense to a legal person, but is not sufficiently clear for a technical person.

So not quite good enough, IMO, but a distinction is being made in the relevant section, and this can be seen if you read the paragraphs in question.

This is not laziness on the lawyers' part, it is them doing what is correct, given the priorities they have been given.

So for some known configuration of orbits of the known planets, there are a limited number of solutions (in terms of mass, inclination, eccentricity, etc.) that you could add to the gravitational interactions of the solar system and still have the current orbits we observe. That gives a reasonable guess about what (another planet) and where (in the sky) to look for to explain anomalies.

But with distant objects that don't emit their own light, even having a good guess of a bunch of the orbital parameters doesn't mean it is easy to find, because you don't have much sunlight reflecting off of it, and the chances it will occlude something else brighter (like a background star) are needle-in-a-haystack level.

Even _if_ you knew the _exact_ orbit you were looking for, there are 360 degrees of sky to search for a tiny, dark object, because you don't necessarily know where in the orbit the object currently is.

Three things immediately spring to mind: scale (1), accountability (2), and profit (3).

1. An automated system can train on data at huge volume, in a way that no single human is capable of doing. Setting aside the issue that training an ML model and artists learning by copying techniques of other artists is, I would argue, fundamentally different acts, _even if we take them to be the same_, we have to acknowledge that in a single human lifetime one person can only "train" on so many works. Automated systems have no such limitation.

2. If an artist violates copyright or oversteps norms around artistic professional practice, they can be held accountable. Companies which violate this by using automated systems so far hide behind those systems ("the AI is doing it/did it") so aren't held responsible (it should be: the company has built the system, and therefore is responsible for how it is used, and what it does). By building up this false sense of agency on the part of systems (which the marketing term "AI" is designed to bolster), lack of accountability is laundered into the actions being taken at scale.

3. Automated systems are, due to their scale, very profitable. I can generate hundreds or thousands of copyright-violating work that dilute the market for artists, and it is incredibly cheap to do so. Fighting those copyright violations in court has to be done more or less on an individual basis (especially if actions like that in the original article continue to fail), which is extremely slow and expensive. If the cost of violating copyright is tiny, and the cost of enforcing it is huge, then it ceases to be a useful tool except for the most well-resourced organizations.

> It seems an ML tool could add a filter to the output and refuse to output a work that too closely resembles one or more work under copyright. Isn't that basically what legitimate professional artists do as well?

No, because copyright is more complicated than "these two things look a lot alike", and legitimate professional artists don't run into this issue, because they aren't constantly trying to skirt the line of "as close as possible to copyright violation while still getting away with it".

> Thousands of artists are capable of infringement, but we don't take away their brushes based on capability.

But they do get sued when they infringe! Enforcement happens, because (for now) it is still possible for independent artists to enforce their copyrights. The argument being made by artists with regard to these ML models is that _they are already infringing copyright_, not that they hypothetically may in the future.

There is extensive literature out there just a google away on the consistent misdiagnosis and under-treatment of women, both as a general trend and for particular diagnoses like heart attack or ADHD. The dismissal of "It's not a woman specific thing" belies the truth that there are women-specific biases against taking self-reported symptoms into account and receiving proper care, and data-collection biases where there simply isn't as much or as sufficient research in how the same disease presents differently in women.

Similar data exists for other groups (for example, pain treatment disparities for black people, because actual doctors with medical degrees think black people have "thicker skin" or "less nerve endings"), but we can acknowledge several kinds of systemic failure of medicine to meet the needs of whole classes of patients without undermining the rightful complains of those individual groups.

[0]: Annual electricity usage from here, using 2018 numbers - https://www.destatis.de/EN/Themes/Society-Environment/Enviro...

Because of differing semantic coverage, a lot of words between German and English are not 1:1 mappings. A reverse example is that in English we often speak of generic entrances, while in German you would normally distinguish between an entrance you drive through ("Einfahrt") versus walk ("Eingang").

[0]: https://en.wikipedia.org/wiki/%C3%9F

The building of it and the budget overrun was a disaster, and demonstrated the worst of infrastructure building in this country, but will happily die on the hill that the result is good, actually. (If we want to get into contenders for "worst in the world", my opening bid is ATL.)

To take an analogy: bad voice recognition software abounds everywhere, not because it is better than what it replaced in terms of UX, but because it works just enough and allows massive cost-savings on hiring people to do customer service jobs.

A world where most marketing copy is written by mediocre AI, and more and more written and visual content are generated by big models that are technically impressive but intellectually hollow is going to be one where the quality of everything sucks just a bit more, but it's so cheap that it becomes pervasive.

(This trend is already apparent and not created by, or limited to, ChatGPT.)

Confidence = voting for the government in confidence motions (or votes which are not explicit confidence votes but considered to be "matters of confidence"). Supply = voting for the budget (because if the government can't pass a budget, then it is considered to not have the confidence of the house, and has to be replaced).

There the goal is less to avoid detection and more to make it hard to determine direction/orientation.

CloudFormation (the AWS-provided IaC product) may have something, though.