Everything in archives pre-2021 is still untainted. All major social media, Q&A, code repos, and archive.org are timestamped. It taints future collection of training data, but not existing collection of training data.
What's the plan then, to coast on pre-2021 data forever? How much utility would todays LLMs have if they were trained on fossilized archives of the internet from 10, 15, 20 years ago?
If you're not interested in random and creative acts of programming, I think you're on the wrong discussion site.
This is for fun. It takes todays headlines and rewrites them. Like I could tell that the "The Mars State Fair Sees Many Galactic Party Candidates but Only One Zorgon" was a rewrite of "The Iowa State Fair Sees Many Republican Party Candidates but Only One Trump".
It retells the present in a future context. I enjoyed it.
Can you please explain why? I care about privacy and security, and this trend seems like we're going backwards. If something needs network access, web apps make sense. But I liked offline tools better when we called them "programs". Web apps to me seem like they turn the web browser into an operating system. I already have an OS.
There are some advantages to using the web as a platform: it’s fully cross platform, completely sandboxed, easy to use, has good backwards compatibility so will likely work forever regardless of OS/browser updates, gives you good debugging tools, and means you can use it without saving it locally if you want
Depending on which OS system you use, your programs can connect to internet and leak your data. As far as I know, neither Windows nor macOS have any built in firewall for outgoing connections. For mac Little Snitch or Lulu are indispensable. A JS app running in the browser sometimes maybe even easier to monitor. You can at least, see the network traffic by the built in devtool.
So this is some kind of attempt to make a Drake Equation [1] for AGI? That's more useful as a thought experiment than something claimed with scientific precision.
> Generally teenage parents won't have very high income, but in the UK this wouldn't be an issue. One of the primary reasons women in my family have kids young is for the financial security. Here in the UK having kids is one of the best paths to moving out of your parents home in your 20s if you're from a working class background.
I'm in the US, where this kind of statement literally baffles my mind, because the opposite is true. In the US, having children is a financial liability and burden. Even for the poorest unemployed single mothers who do qualify for government benefits, it's a net financial loss to have a child. Our food assistance programs are pathetic and the shame of the developed world. Childcare is unaffordable even for entry-level tech workers these days. Students who have a child during high/secondary school graduate at 10% of the rate of those who don't.
> I guess could someone just give a good argument against teenage pregnancy? Ideally one argument that doesn't rely on your subjective values about what a fulfilling life involves?
Any question about the role of government (especially around family planning) ultimately is a question about what it means to have a society that supports living the good life. So for millennia, we have been arguing about what the good life is. A good answer to your question must recognize that in a society, we have to come to a consensus around subjective values about what a fulfilling live involves, in order to make rational decisions about how to further those agreed-upon ends.
If you try using an 'objective' metric to avoid making decisions based on subjective values, then you're just not noticing the subjective values that are correlated with using that metric over others. The school graduation stat I cited earlier has a value built in: it assumes that all other things equal, it is a bad thing for students to not complete secondary school, because it gives the foundation to be a good member of a good society.
> I'm in the US, where this kind of statement literally baffles my mind, because the opposite is true. In the US, having children is a financial liability and burden.
The UK is interesting. Most of my middle class colleagues can't really afford children so if they have them will have them later in life and maybe have one or two. But my working class family and school friends have loads of kids, and this is fairly normal here. The affordability issues of having kids exists here, but only if you're working. If you don't work having a kid fast tracks you to council housing and gives you extra disposable income. Another "hack" most of my family do is then to get their kids diagnosed with behavioural issues and then you'll "make" another ~£1,000/month tax free.
What I'm saying here though is quite controversial. A lot of middle-class people will accuse me of lying and suggest that working class people really want to work. They do not accept people choose to go on benefits because living in a £350,00 home and not having to work is a better lifestyle for the majority of people from my background because a £15,000/year job isn't even going to get you a small flat here.
> Any question about the role of government (especially around family planning) ultimately is a question about what it means to have a society that supports living the good life. So for millennia, we have been arguing about what the good life is. A good answer to your question must recognize that in a society, we have to come to a consensus around subjective values about what a fulfilling live involves, in order to make rational decisions about how to further those agreed-upon ends.
I think my issue with teenage pregnancy is that it's used as a heuristic based on middle-class assumptions about what a good life involves and when a responsible individual should become a parent. Perhaps part of the reason I have a different view here is because for most women in my family being a mum is so important and going to college and pursuing some grand career is largely a waste of time when you couldn't even get decent grades in school.
If I had a child and they had a kid young I'd be a little disappointed I think because I'd like my child to do something more novel with their life which generally requires a good education. But this is my middle class aspirations speaking. A lot of people rightly value family over their career I think. This is perhaps especially true of women.
My interpretation is that it gives a false sense of how difficult it actually is to raise a baby.
First, they only had the students care for the simulated baby for five days, and 'childcare' was provided for free during their school day (they had to turn it off during the school day). Five days is not long enough for the novelty to wear off and the day-to-day reality to kick in.
Second, the dolls are not a good simulation of what it is actually like to care for a baby. It's a step up from a tamagotchi. Changing a diaper on a real baby is nothing like changing a diaper on one of these dolls. Real babies scream and cry much louder, and often don't stop even if you do everything right.
tl;dr: Students at schools in both the control and intervention groups took the standard required health class, which includes traditional sex ed. Students at schools in the intervention group had to opt in to the baby simulator. It wasn't a regular class and was independent of the required health class, but it did take place at the school. It was one week, 40 minutes a day to learn about the baby simulator and a few other basics. Students at schools in the intervention group only cared for the baby for 64 hours: 4 days x 16 hours. 'Free childcare' was mandated for the 8 hours they were at school, as they were turned off during the school day.
> Individual participants were not blinded to their group assignment and upon giving consent were aware whether they were participating in the VIP intervention or control arm of the Trial.
So my read of the results is that participants who were interested in having a child (and were in an intervention school) selected into the VIP program. Participants in the control schools were a more random group.
No thank you. That's not a unique benefit of the patent system, that's a minor ancillary benefit that has already been replicated elsewhere. The patent office doesn't have a patent on historical records of so-called inventions.
And the patent process is not optimized to produce documents that actually help other inventors or future historians. In fact, to the contrary, as it exists within a particular narrow legal IP regime that optimizes around legal risk. This is especially the case with software patents, where the patent office incentivizes patents that use convoluted language to make an obvious process seem like a novel invention.
In other words, let's assume a future where there are no more copies of Zelda to play. If you had to choose between getting to preserve the patent on the Zelda loading screen and an actual recording of it, I'll take the recording every time. Or in a future where much of the knowledge of computer science and programming was lost, I'd rather have an archive of a set of textbooks, Stack Overflow, and Github than the entirety of the patent office's software patent documents.
I think your post reflects several misunderstands or false assumptions about the patent system.
First, you misstate the purpose of disclosure. It isn't to become a record of historical inventions, it's to encourage inventors to disclose their innovations, as inventors otherwise would not disclose their inventions. Inventors and their businesses would instead be incentivized to wrap their inventions up in trade secret and never disclose anything at all.
Second, patents don't use convoluted language to hide an obvious process. That's (a) merely facial and a waste of time, (b) contrary to the actual legal goal of patents which is to encompass as much in your patent as possible, while still maintaining its ability to grant. Patentees must actually disclose their invention, in a way that's cognizable to someone skilled in the art, or they simply do not have any benefit from the patent at all.
Inventors and their businesses would instead be incentivized to wrap their inventions up in trade secret and never disclose anything at all.
In the past there was perhaps more benefit from this exchange: manufacturers disclosed special knowledge of their products in patents and the world at large gained knowledge that would otherwise remain secret. But this theoretical exchange breaks down if the secret is easily reverse engineered or otherwise unlikely to remain secret. Reverse engineering all manner of products is much easier now than it was 100 years ago but the duration of patents has remained fixed at 20 years. I doubt that it would take more than a year to reverse engineer these patented Zelda features (or how to make a new small molecule drug, for that matter) in the absence of disclosure through patents.
There's still an argument for patents to incentivize investment in R&D. We still want people to invest the time and effort in developing and proving new small molecule drugs even though modern instrumentation and synthesis planning makes it easy to copy a drug. But the "disclosure is better than secrets" argument in favor of patents has been weakening every year as secret-keeping becomes harder.
