HN user
The correct way to do it would be to force Anthropic to remove content and the results of training based on that content from their models at copyright holder's request.
aka. "It's not our problem now."
Brilliant. This reminds me of (but is way more advanced) a single-file Python NoSQL key-value DB that used pickling to store data. It was FAST. Can't remember what happened to it. Anybody?
What a wonderfully visual example of the crap LLMs turn everything into. I am eagerly awaiting the collapse of the LLM bubble. JetBrains added this crap to their otherwise fine series of IDEs and now I have to keep removing randomly inserted import statements and keep fixing hallucinated names of functions suggested instead of the names of functions that I have already defined in the same file. Lack of determinism where we expect it (most of the things we do, tbh) is creating more problems than it is solving.
A few years in corporate managerial role will turn you into a bot.
I used to work with manager who told us to "have balls, deploy straight to prod" We spend a month cleaning this up. The other time when another manager demanded I "run this script" from the iphone he shoved in my face a whole warehouse went down in Ireland a month before Christmas and the Police were involved. There is an irrational push to deploy AI everywhere for no reason whatsoever. It, well the LLM part of it, is bad tech and all those "AI success stories" are pure hot air.
A similar thing is going on at Amazon. Their AI is rewriting perfectly good product descriptions provided by manufacturers and I noticed that I stopped buying there, because I cannot find out what it is that I am buying or how much of it I am buying. I started shopping at sites that sell same products, but do not change manufacturers' descriptions. I have a feeling that post AI bubble we'll be doing a lot of manual cleanup of information "improved" by AI.
I'd love to see Dennis Ritchie commemorated as one of the greatest Americans.
Can you not install Parallels and use it to run Linux under Mac OS X?
I wish they published CPU and graphic performance comparisons vs M1. Apart from coding I do a lot of large photo (100MP) and video (4K+) file processing and I am interested in core graphics performance. Talking centipedes or flying trains are not my thing so their AI benchmarks are useless to me.
Props! I hope they keep it and don't use it as a play to get a better deal from a commercial provider. I am jaded after seeing too many "digital transformation" projects running on a 3-5 year cycle of switching from Offie 365 to Google then back to Office 365.
A remnant of their 'Linux is a cancer' phase and the famous fight against Open Source that started with the "Halloween Document" then helping SCO https://yro.slashdot.org/story/21/04/01/1647259/sco-linux-fu... They are not Open Source's friend and at one point extracted license fees from Amazon for using Linux... https://www.zdnet.com/article/amazon-becomes-the-latest-comp... Yeah, I wish they would leave OS alone.
Not as much as it used to be. Before cloud computing became a thing, if you wanted to squeeze the last bit of performance out of hardware, FreeBSD was the way to go. Yahoo! used it when Yahoo! was the biggest site on the internet. Over time Linux became more performant and ever since it has become the OS of choice for AWS and other cloud provides, FreeBSD's popularity has dropped.
Given that governments want to move to 100% digital government and economy, denying someone access to the internet will strip them of ability to take part in society, use government services (accessible via digital id), or pay (living expenses, travel, repay debt, etc.)
Find at least 10 things wrong in that video...
The solution is to ban drone deliveries in built-up areas. We do not need them.
or Kinder Surprise ;-)
Estonia (and now Ukraine) have worked on being able to do a "backup" of the country and a "restore" elsewhere if needed. (I am oversimplifying, but contingency plans have been part of the overall design that eID is a part of.) The UK doesn't have such designs and contingencies in place. The private sector is no better, every year there are major security breaches. It is premature to stick Digital ID onto a rickety network of badly secured databases.
It's interesting how an idea almost nobody wants to see implemented in practice keeps coming back thanks to the efforts of the person most of the country, including his own party would rather forget.
Amazon are no longer the golden standard of e-commerce. I think 5-10 years from today we're going to look back at 2025 as the year Amazon started to destroy itself from within. They are pushing AI to "update" and "optimize" product descriptions. It's already made art supply descriptions a mess and now I see the same thing happening in the music gear section. I noticed that I go to other sites to buy stuff I was planning to buy on Amazon, because I am not sure what I'm buying anymore on Amazon.
I started building on Cloudflare, but after their "pay us 120k or else" tactics they got famous for I decided to move code elsewhere.
Managers on AI projects need to show "mass adoption" in order to survive the next annual review. That's why you have pointless AI stuffed into every orifice of Google/YouTube.
That's a lot of words to make oneself feel better about manipulating others.
- There is no doubt a large volume of abuse by tech consulting companies. It's likely even worse than it looks, because the H1Bs in the U.S. are to support even larger teams offshore. I don't understand why we can't just blacklist these companies.
I wish we could ban offshore IT consultancies dumping "talent". Unfortunately politicians will not support such ban, because those consultancies help corporate donors keep the wages down. If the government banned suppliers who use offshore consultancies from bidding for government contracts those outfits would disappear overnight.
AI Santa next?
Microsoft and nuclear reactors... what could possibly go wrong?
This will paid to rights holders, not authors. Published authors sign away the rights to financial exploitation of their books under the terms of contracts offered. I expect some authors suing publishers in turn. This has happened before when authors realised that they were not getting paid royalties on sales of ebooks.
AI is the biggest threat to bs middle management/ba jobs in recent history. Those people panic and try to become "AI ambassadors/advocates" to survive. They are pushing the narrative so that they have answers and "metrics" of adoption to show to upper management and survive another round of layoffs. You can see how it works at YouTube, where someone decided for no good reason to "upscale" Shorts making them all look bad. This is done automatically, without asking for the creators' permission and you cannot turn it off. The results are crap, but who ever made that decision can brag about widespread adoption of AI and survive another annual review.
Out of curiosity, would CouchDB not be a contender for this type of application?
The US government used to consult sci-fi authors, especially when they were writing stories somewhat based on extrapolations of current advancements of technology or science. I'm not sure any of the notes from those meetings are available online, but I'd love to read them and compare.