kevingadd

Not that weird of a strawman:

> According to the Metro, several states have prosecuted women for miscarriages or stillbirths. They include South Carolina, Georgia, Arkansas, Texas, California, Mississippi, and Ohio. It appears West Virginia has joined the ranks as well.

> Though California has since passed laws banning criminal charges and investigations of pregnancy loss, it previously jailed two women for stillbirths.

Also:

> Raleigh County Prosecuting Attorney Tom Truman told WVNS 59News that a number of criminal charges under state code, including felonies, could be levied against a woman who flushes fetal remains, buries them, or otherwise disposes of remains following an involuntary abortion, also called a miscarriage.

We've established arbitrary lines on when the embryo progresses to the point of having a legally mandated funeral in some parts of the US.

https://www.ndtv.com/world-news/new-law-in-us-state-could-le...

On this note, even if LLMs or "AGI" make businesses more profitable by improving margins and allowing them to cut labor costs, you eventually have to balance it out somehow - all those people who've lost jobs are competing for a smaller number of available jobs, and the people who can't find jobs will either need to be supported by taxpayer dollars or die in poverty.

People talk about UBI as a solution for this but the money for UBI has to come from somewhere. Are corporations going to voluntarily pay a new tax to support UBI?

It's interesting that Valve sort of put themselves in this situation by opting not to police their store anymore.

I'm personally a fan of fewer restrictions on content in video games and fewer "gatekeepers" but it's kind of inevitable that people would get upset when you chose to allow people to sell games like "Sex With Hitler" and "Pimp Life: Sex Simulator". Deciding to allow that content on your store and simultaneously not going to bat for it is weird, it's like they decided to just get the porn money while they could as a short-term boost to revenue.

Itch.io still has fewer restrictions but I assume they'll eventually have to clamp down too once payment processors cut them off - they don't have the financial resources to fight it like Valve or Epic do.

Interestingly Nintendo has as of late relaxed their restrictions too, you can find porn-adjacent shovelware on the Switch eShop despite their history of being very censorious. I wonder if payment processors will successfully push them around too or if Nintendo is too big to get pushed around.

My understanding from reading the federal website is that this would not be Medicaid, it would be a different program. So they would not be in the Medicaid database, right?

This is not what the federal website for Medicaid says, though.

I don't understand why ICE would need access to Medicaid data. You need to be a citizen or lawful permanent resident to access that program, not to mention all the other additional criteria. The idea of illegal immigrants somehow bypassing all the checks and balances successfully en masse feels a little silly to me.

Just a quick check of the official website to try and get onto Medicaid in WA state shows that it requires a social security number and citizenship information: https://www.wahealthplanfinder.org/us/en/health-coverage/get...

IIRC the free tax filing thing being killed is just classic lobbying, the owners of Intuit and other paid tax filing companies have myriad political connections and spend a lot of money to protect their interests.

Does this work in practice? Historically my actual residence is needed for tax reasons, i.e. properly tracking what state I owe income tax in for cash compensation and stock options.

Beyond that, as we gradually slide towards having zero political representation and near-zero actual human rights, it becomes more and more questionable what we're getting for our tax dollars. The contract between the taxpayer and the government is being eroded.

This post gradually seems to tiptoe towards eugenics, which makes me a little nervous, closing with this bit:

> If we took this history seriously, we might spend more money on not only parents of young children but also the basic scientific breakthroughs that would make it easier for future parents to have the children they want, whenever they want them.

This is in the context of enabling broader fertility by making it easier to get pregnant, to be completely fair. But for me it does raise the question of what 'the children they want' looks like in a modern climate where heritable traits not only affect your capabilities in life but now dramatically impact how you are treated, whether it's being mistreated based on skin color or being at a disadvantage in education & the workplace due to conditions like adhd, chronic fatigue, etc. Raising a child with heritable conditions (or random genomic quirks) can also be much more expensive than a child that is closer to the norm, too.

I'm still not sure where I land on the question of whether it's appropriate to try and edit these 'disadvantageous traits' out of an embryo. It seems like a classic slippery slope problem and I don't know if it's possible to trust anyone (or anything, if one were to suggest AI as a solution) to navigate it right.

Off the top of my head:

If you rely on being able to do things like mark a range of memory as read-only or executable, you now have to care about page sizes. If your code is still assuming 4KB pages you may try to change the protection of a subset of a page and it will either fail to do what you want or change way too much. In both cases weird failures will result.

It also can have performance consequences. For example, if before you were making a lot of 3.5KB allocations using mmap, the wastage involved in allocating a 4KB page for each one might not have been too bad. But now those 3.5KB allocations will eat a whole 16KB page, making your app waste a lot of memory. Ideally most applications aren't using mmap directly for this sort of thing though. I could imagine it making life harder for the authors of JIT compilers.

Some algorithms also take advantage of the page size to do addressing tricks. For example, if you know your page size is 4KB, the addresses '3' and '4091' both can be known to have the same protection flags applied (R/W/X) and be the same kind of memory (mmap'd file on disk, shared memory segment, mapped memory from a GPU, etc.) This would allow any tables tracking information like that to only have 4KB granularity and make the tables much smaller. So that sort of trick needs to know the page size too.

Are you suggesting that YC applications contain trick questions? Or that the questions are intentionally drafted in a confusing and/or misleading way? That seems kind of weird.

The CP2077 framerate is 59, so it's probably effectively running at 60FPS, or is CPU-limited instead of GPU-limited. If the graphics side of things has spare headroom it's going to draw less power.

On the flip side, the CPU may be doing less work due to things like more efficient drivers / less OS nonsense tying up CPU, which would also reduce power draw.

My understanding is that chaining compressors is a classic technique for image compression. IIRC PNG and Basis are both implemented as initial transformation/pre-filtering/conditioning pass(es) designed to make the image data more compressible before feeding it to a codec like gzip or zstd.

This definitely works for things that aren't images too. I previously proved that you could improve the compression ratio for WebAssembly significantly by performing lossless transforms on the module before feeding it to gzip or brotli (though the gains are much smaller for brotli since it's so good to begin with): https://github.com/WebAssembly/design/issues/1180

Mesh optimizer's performance here is a nice reminder: the state of the art in general purpose compression is hard to beat, but special purpose still has room for improvement.

The Hertz from this article is a company you should never do business with, unless you like getting falsely accused of a felony, arrested, and imprisoned for no reason:

https://www.npr.org/2022/12/06/1140998674/hertz-false-accusa...

I personally use other companies when I rent a car... but I'm also paranoid about the kind of nonsense described in the article, so I just buy the damage coverage when I rent the car. It means renting a car is far more expensive than it should be, but it saved my bacon the last time I rented when someone dinged the paint.

It's funny that the AI bogeyman is used in the article to make this new tech seem scarier. The problem is the general practice of trying to milk you for money over regular wear and tear.

This is presently happening in China with homosexual literature: https://www.bbc.com/news/articles/c056nle2drno

Rural police departments decide that a piece of text 'harmed' one of their residents and prosecute the author.

How will you sell them porn when VISA and Mastercard ban you?

If this screenshot isn't omitting some truly exculpatory context (I can't imagine what kind of context would justify it), it appears to be Grok advocating for truly abhorrent behavior: https://bsky.app/profile/kthorjensen.bsky.social/post/3lti7l...

Though "advocating" is probably too anthropomorphizing, I'm not sure what the right verb is for this.

My personal experience: We run over 1.1m test cases to verify every PR that I submit, and there are more test cases that don't get run on every commit and instead get run daily or on-demand.

At that scale getting quick turnaround is a difficult infrastructure problem, especially if you have individual tests that take multiple seconds or suites that take multiple minutes (we do, and it's hard to actually pull the execution time down on all of them).

I've never personally heard "we don't have the budget" or "we don't have enough machines" as answers for why our CI turnaround isn't 5 minutes, and it doesn't seem to me like the answer is just doubling the core count in every situation.

The scenario I work on daily (a custom multi-platform runtime with its own standard library) does by necessity mean that builds and testing are fairly complex though. I wouldn't be surprised if your assertion (just throw more resources at it) holds for more straightforward apps.

Google did it the legal way with Google Books, didn't they?

If most of what an LLM spits out is a digested version of its training set, is it really an outside view of the world? If anything, seeing how easy it is to get these things to spit out conspiracy theories or bigotry suggests to me that we're far from being able to get a robot's view of the world.

Though for some people if the "robot" says bigoted things or supports their conspiracy theory of choice that's just "proof" that their viewpoint is correct. Tricky to navigate that problem.

The article gestures at it:

> Even before these recent changes, Grok raised eyebrows after appearing to briefly censor unflattering mentions of Musk and his then-ally President Donald Trump, repeatedly bringing up “white genocide” without prompting, and expressing skepticism about the number of Jews killed in the Holocaust.

But I'll be specific: Between them obviously rigging the system prompt to push it towards a certain answer on "white genocide" subjects in the past, the strange obsession with Jewish people in Hollywood suggests an unusual training set at best.

Putting your thumb on the scale so obviously in any direction feels really questionable if your goal is to turn your AI product into a popular, profitable thing. But maybe that's not xAI leadership's goal at all and they're happy to just light money on fire to satisfy some particular egos by making sure the answers to key questions are as desired, regardless of what a normal training set would otherwise generate.

Can't generalize based on one data point but it really was historically low-stress and low-toxicity in every part of the company I've interacted with in the past 8+ years. If it had been combative and high-stress like some of the past places I've worked I wouldn't have considered the inferior compensation acceptable.

We've had a considerable number of layoffs impact the video game industry over the past few years ( https://en.wikipedia.org/wiki/2022%E2%80%932025_video_game_i... ) and executive incompetence is a big factor in it.

Deciding to set a "success" target of 100 million players and then spending upwards of $400 million USD developing one title is a recipe for studio closures and/or layoffs when it inevitably "fails" because executive leadership didn't set reasonable targets or come up with a reasonable budget. It's a big house of cards.

A more diversified portfolio of titles with more reasonable budgets would be a much safer choice, and it's how things were done successfully in the past.

The baseline number of attacks was in the single digits, yes.

Unfortunately a lot of modern software triggers UAC popups now. Games (for anticheat and/or network connectivity), development tools (for network connectivity or debugging), updaters for stuff that live-updates like Electron apps, etc.

This seems like a liability nightmare. If they're just scanning all the image files on people's devices and using them for training, they're inevitably going to scoop up nudes without permission, not to mention the occasional CSAM or gore photo, right? Why would you want to risk having stuff like that sneak into your training set when you already have access to all people's public photos?

> Commuters in the Philadelphia area are now set to lose access to 50 bus routes, 66 rail stations, five regional rail lines and Sports Express trains. Service will be cut by 20% starting August 24, with a further 25% reduction on January 1, 2026. In September, a full hiring freeze will go into effect and fares will rise to $2.90 from $2.50.

> The downsizing is a stark illustration of the fiscal reckoning that mass transit systems in some of the nation’s major cities are facing as they deal with a collective $6 billion funding shortfall. Agencies in Chicago, Boston and San Francisco have been scrambling to raise funds for their operations before the start of the next fiscal year on July 1.

It does seem right that the article refers to a "death spiral" here. Once transit becomes insufficient to get from place to place you stop using it and pivot to cars, and then there's even less justification for funding the transit system.

I wonder if American society will ever come up with a solution for this since the problem is so widespread. More cars on the road doesn't seem sustainable, traffic is already really bad in many places.