I am unable to extract any meaning from your post. You appear to be making a general claim: it is impossible to design a programming language where everything is a value. You at least admit that "data thingies" can be values. Are you claiming that it is not possible for functions to be values? (If we assume that the argument and the result of a function call is a value, then this would mean higher order functions are impossible, for example.) If not that, then what? Please give a specific example of something that can never be a value in any programming language that I care to design.
I see this posted everywhere this week. Is it really that good? I understand this runs on any hardware (not limited to Mac Minis) as long as you have an API key to an LLM (Preferably to Claude).
It sounds interesting to me, I might install it on a cheap Mini PC with Ubuntu.
You say that as if it dismisses any reason for non-visible traits to diverge, but it's commonly theorised that this is actually the main reason. Taking early Europeans, for example, there are very clear and extreme selection pressures introduced by the long, barren, freezing, dark winters.
Come to think of it, even without these pressures it would be highly unlikely, in the Out-Of-Africa model. What are the chances that each and every small group migrating to each corner of the globe all across time all happen to be a perfectly average representation of their larger former tribe? It's absurd to suggest that the few individuals to first settle Australia must have been completely identical in every non-visible way to the few individuals that first settled Iceland.
If you selected two random groups of 100 people from the same 'tribe' today, and managed to quantify various mental and intellectual attributes, you would certainly find variation between them! Then put these two groups in radically different environments for thousands of generations...
>not only is there not extraordinary evidence for it, but the evidence we have is mostly countervailing
A lot of people try to explain away correlations between race and IQ, criminality, wealth, relationship statistics, etc, but I've never seen someone suggest that these phenomena just don't exist. I'm not sure how I'm supposed to respond.
See also John Barnes' Century Next Door books, where "memes" are basically computer viruses that jump to running on human brains, not just silicon chips. The results are... not pretty.
Most of the social spaces that I frequent don't have the amount of political topics posted as HN.
Would you like to know the difference between those spaces and here? It's that in those spaces, regardless of if the members are left right or center, the community is on the same page in terms of authoritarians, and authoritarian apologia will get you tossed.
Therefore, there isn't the same sort of desire - or need - to point out the obvious and show the uncomfortable realities to the crowd.
Refusing to take a stand on this sort of thing and leaving it for the community to sort out will only make things worse. It's functionally no different than the kind of combative environment you get on major social media networks.
I think the article talks about nonviolent protests - the first two were anything but.
The Slovakian incident worked, because Slovakia has a working representative democracy.
In a deeply flawed, or downright nondemocratic system, like Serbia or Georgia, it's very hard to drive change through nonviolent protests.
It also bears mentioning, that the key issue with protesting, is that it, legally speaking does nothing. Legal representatives are under no obligation to do anything in response to protests.
The left / right split isn't really meaningful in the United States right now.
The split is currently between people who believe in and want a functional and equitable government, and those who are fine with a kleptocracy as long as they are personally the beneficiaries (or at least, the people they dislike suffer worse).
People like Frum were quick to notice this and get on the correct side of it. Unfortunately, there are not enough Republicans who feel the same way to make much of a difference.
I do think it's an interesting quirk: why does the model stick to the incorrect answer? I suspect part of the reason is that it's "reasoning" has the same flaw: the numbers are chosen because they look likely and not because they are correct.
> Storing prompts and contexts along with generated code is likely how we are going to be doing software engineering for a few decades
You specifically asserted that prompts would be the future of software engineering. If my memory is not mistaken, this was edited later to hedge with "likely" and did not include that when originally written?
It is in a sense by design because the focus was creating a decentralize-able/federate-able protocol and infrastructure that can scale more or less indefinitely first and foremost, community second.
The community is working on actually decentralising the network now that things mostly "just work" (assuming you are using did:plc/generally a happy path user).
- Building out PDS communities that are trusted takes time and nowadays there's a few outside of bluesky PBC (one or two big ones and a bunch of smaller ones). People are eager to move off because a lot of users really really don't like bluesky PBC leadership but it's a matter of waiting for these third party communities to reach critical mass.
- Relay infra is already pretty much decentralised. Lots of people still rely on the main relay but it's trivial to use a third party relay and there's more of them than you can count.
- There are a lot of really high quality third party clients and afaict a lot of users do actually use third party clients but there's basically no metric for tracking these stats.
- Appviews are expensive currently and there's work on making them easier to host but there's already one "full" alternative appview for bluesky.
- There are a lot non-bluesky apps/services that are genuinely high quality experiences and they are gaining their own communities.
The main technical barrier to true decentralisation outside of improving UX is introducing other did:methods and/or spreading trust of did:plc across the community (ex: clustered via raft or paxos across major operators) but there's just not a reason to pursue this over the other fires that need fighting in the ecosystem right now (and keeping did diversity low reduces another source of complexity the space just doesn't need to tackle yet).
--------------
TLDR: it is intentional because the goal is to in order of priorities:
1. get the architecture for eventual decentralisation right.
2. make it exist.
3. make it good.
4. make it easy to use for normal people.
5. build community.
6. focus on decentralisation.
Decentralisation in theory is the first priority but in practice it's the last priority. Being able to decentralise is always the utmost importance but forcing it to happen is not ever the top priority because that's on the community, not on the developers.
There is a 15-30% difference between the groups at baseline (fig 8c-9c, 8d-9d), about the same magnitude as the claimed effect of the experimental condition.
I think the result would be much stronger if these baselines were comparable, so they show they have accounted for other variables like time of day and light history. I am also skeptical of any effect in the retina lasting 6 weeks, with no fading.
Consider that people are often exposed to much more infrared light outdoors, so "worked under a relatively dim incandescent lamp" is not a particularly novel stimulus. Imagine that any of these people spent time outdoors during the six weeks - thousands of times more infrared light there.
If it was actually a terrible place the illegal immigrants would leave on their own volition and it wouldn't be necessary to have federal police find them and forcibly arrest and deport them.
And comes with a price tag paid to people who neither own nor generated that content. You don't think that shifts the ethical boundaries _significantly_?
This has nothing to do with immigration law. If it did, there would be no offer on the table to withdraw the ICE troops in exchange for the MN voter database.
I've been dismayed by how fast the "we should own our hardware" crowd has so quickly radicalized into "all security features are evil", and "no security features should exist for anyone".
Not just "there should be some phone brands that cater to me", but "all phone brands, including the most mainstream, should cater to me, because everyone on earth cares more about 'owning their hardware' than evil maid attack prevention, Cellebrite government surveillance, theft deterrence, accessing their family photos if they forget their password, revocable code-signing with malware checks so they don't get RATs spying on their webcam, etc, and if they don't care about 'owning their hardware' more than that, they are wrong".
There's no uncertainty. Republicans now openly assert the 2nd amendment belongs to supporters and defenders of the regime, and no one else.
The movement opposes equality because equality stands opposed to their need for hierarchy. It is a domination and submission movement. It boasts about its application of double standards. Double standards are not logical fallacies, when they use them they are virtues. To enjoy for themselves what they deny to others is a display of dominance.
There might be Gandhian/Nelson Mandela way of handling this.
Both fought to change system and didn't teach hatredness towards individual a Racist person.
Get arrested peacefully.
Others can work for immediate protest for release of the arrested.
Appeal to common values that they also have, and show how they are violating the religious values they profess.
Technically someone can make some app, that can easily help in getting the citizenship proof for an individual.
I have implemented similar behavior in some of my projects. For one, I also have also implemented 'cursors' that point to some part of a value bound to a variable and allow you to change that part of the value of the variable. I have used this to implement program transformations on abstract parse (syntax) trees [1]. I also have implemented a dictionary based on a tree where only part of the tree is modified that needs to be modified [2]. I have also started working on a language that is based on this, but also attempts to add references with defined behavior [3].
I did not suggest that; however, the law is clear. If I use my knowledge to produce code, under a specific license, then you take that code, and reproduce it without the license, you have broken the law.
You can twist this around as much as you like but there are several studies showing that LLMs and and will happily reproduce content from their training data.
ICE officials are pretty consistently saying that they do more visible immigration enforcement in places where the local police are forbidden by local or state law from giving information about people they arrest to ICE, compared to places where the local police do this happily. Legally-forbidding local police from cooperating with federal immigration enforcement is a prototypically blue-state policy that red states do not generally do.
The visible disruptive protests against ICE activity are also the sort of thing that you'd expect the sorts of voters that make a blue state blue to do, so when ICE does arrest illegal immigrants in red states, there's much fewer people who are inclined to protest it and therefore less publicity in general.
