Just wait until all the conversations you've ever had with AI (which 100% is training on them as well as keeping it's own memories about you that you have no control over) starts getting used to answer questions other people have asked about you.
That's my theory of what's to come, anyway.
People talk to these things not understanding the implications, and can get extremely personal. The model and companies behind it know who you are, you discuss details that reveal what you do, where you live, where you work, what you search for, and you probably signed in with an oauth provider like github or google, which is more than enough of a thread to start pulling on to learn more about you/link other things to you from on the open internet. It'll all get sucked up into the model and before you know it I'll be able to ask a model about my coworker (you) and get back answers from conversations you had with a model a year or two prior, exposing details about you that you might not want out there. And even if that isn't supposed to be allowed, how well has it worked out so far when it comes to data exfiltration and guardrails. If the model has info on you, being told not to share it won't protect you or that data.
Some people like messing with cars. They take the time to understand what's happening and learn the process and pitfalls. Hobbyists wiil never be as good as trained professionally but we can still get the job done. I went through the trouble to diagnose and replace a bad alternator on my civic after the battery started dying too fast. I did it cause it was fun.
The other reason i did it is because the dealership and other shops quoted me over 10 times the cost of parts, and I literally did not have the money to take them up should i have wanted to. Car maintenance is expensive, _especially_ at the dealership.
> projections of what the world might look like when this happens
I've done this a few times. A world with 0 privacy would definitely be safe (given benign governance), but also would likely be pretty boring. Crime would become a non-issue as everything about everyone being easily known/knowable by everyone else means the root of any given crime, some desire/need, could be brought to the fore and resolved before it became an actual issue. But also there would no longer be any kind of surprise in anything; everything and everyone would essentially become dull and grey, and humanity isn't about that kind of life experience at all.
I don't really understand this mindset that being at home and raising your kids is only something you do when forced to. For my family, if we had more options -- ie, more money -- then both of us would be stay-at-home parents. It's much more of a joy than going to work.
Oh, and good chance that in the summer months (when demand is pretty high, especially in the South of France) they may have to shut down again because of a lack of cooling capacity. France was ahead of the game in the 70's, but should have invested a lot more than they did since then. That they installed more than they needed also didn't help, especially not because the energy produced is sold on the open market at a net loss just to keep the reactors operating.
And last but not least: they have an ever growing waste problem.
The cost of which (besides the maintenance costs mentioned above) has not been accounted for in the electricity pricing. If they did, they'd be running at an even higher loss. Probably the next generation will be presented the bill for that.
Interesting timing, because I think it was in March when I had a chat with Gemini about what the heck these things are supposed to be called, and that's where I first heard the term.
It's probably just a coincidence. But that would be pretty interesting if we have an example of some kind of memetic phenomenon where one or more popular LLMs makes a claim that people then start to repeat as true, or at least follow up on it and start writing about it, and in so doing the claim becomes true. Even if it didn't happen in this case, I feel like it's only a matter of time.
That map doesn't effectively capture the intermittency of solar energy in different climates. In Britain the country gets less than 8 hours of energy during peak winter. It also often goes with overcast skies for extended periods of time. A bigger array does not solve these extended periods of non-production.
I fed it my most-read blog post and asked it to identify me and it confidently asserted it was written by Kelsey Piper. Maybe some writers just take outsized importance in Opus' "mind".
Opus 4.7's initial bomb (its actually not that bad) was the best thing for my wallet since my wife left me for my neighbor.
I have the $100 monthly subscription and ive gone on a 36 hour coding tear and havent even checked my usage. Probably should...
Here's my latest 90% build - might finish might never open the directory again --- anyway the idea is codex is integrated via app-server which utilizes the chatgpt subscription model and the Codex CLI. It auto connects if you have Codex CLI and is ready to go from day one with full integrated custom deterministic javascript and Codex manual tools mainly for YouTube and Reddit extraction and its pretty good.
Basically its computer use in the browser with local enviroment access. The chromium browser and chat are intergrated into an Electron Desktop app which allows massive opportunity for integrating codex as far down the chromium tree as possible. Doesnt have the network tools yet or a few of the big boys but this is still DEV. Check it out ----
https://github.com/chillysbabybackribs/Goldenboy-YouTube-Red...
If you're referring to the article's description of the study's measure of success, the metrics had little to nothing to do with direct financial gain.
> [...] in the sense that they have vanishing influence on public discourse.
Nothing inherently requires art to be a part of the public discourse. Sometimes artists create art for art's sake, and/or just to make a buck. Sure, occasionally some art makes it big in the public eye and becomes part of the zeitgeist, but the vast majority of art barely sees the light of day.
> people within the species would not easily recognize it
[Citation Needed]
Sorry, but I had to. There's easy counterexamples of true, species-wide biases that we're fully aware of. Optical illusions, cognitive biases, cultural universals (community-sanctioned relationships/marriage, inheritance, ceremonial treatment of the dead). What we don't have are universal biases towards believing specific facts or stories.
This is neat, but I will stick with Instax wide. With a $1000 mint body you can get full control of the film. Is it the same aspect ratio? No. But I can get film at Target and it’s instant. Very cool, any analog film is awesome, but this price just isn’t sustainable.
I have one of those and can attest to its quietness, as well as reliability --- it's almost 2 decades old and still working well with no sound, just needs an occasional cleaning. Takes almost 30 seconds to coast to a stop!
I've been using qwen 3.6 with oMLX on my M1 Mac Studio and it's been awesome. Took a while to get things set up, figure out which of the hundreds of models would be a good fit for my use case, and then get it strapped into opencode's harness, but it works! Its slower than a hosted model, obviously, but I'm tickled pink that I can give it a relatively complex chore, like I would've with my a Claude Pro subscription, and it'll churn away on it with good results and no god damn arbitrary usage limits.
Oh you're right to push back. I just love saying this nowadays :P Anyway, I haven't used these languages in a long time but the code looked like php to me, though I did notice the .pm file extension and wondered where I've seen it before.
Oh, that's what the article was referring to with "interpolation".
Weird that I didn't hear about it before, it's not that used in practice?
One reason I could see is that binary search is fast enough and easy to implement. Even on largest datasets it's still just a few tens of loop iterations.
More like a few decades, if ever. Battery production is estimated to reach 6.8 TWh per year in 2035 [1]. But only 12% of this is expected to go to battery storage systems, yielding just 800 GWh. 12 hours of storage worldwide at current electricity demand is 30,000 GWh. And by 2035, electricity demand is going to be more than that, as transportation and industry is increasingly electrified and migrated off of fossil fuels.
And again, this is just for 12 hours of storage. Seasonal fluctuations can depress renewable generation for days or even weeks.
That's my theory of what's to come, anyway.
People talk to these things not understanding the implications, and can get extremely personal. The model and companies behind it know who you are, you discuss details that reveal what you do, where you live, where you work, what you search for, and you probably signed in with an oauth provider like github or google, which is more than enough of a thread to start pulling on to learn more about you/link other things to you from on the open internet. It'll all get sucked up into the model and before you know it I'll be able to ask a model about my coworker (you) and get back answers from conversations you had with a model a year or two prior, exposing details about you that you might not want out there. And even if that isn't supposed to be allowed, how well has it worked out so far when it comes to data exfiltration and guardrails. If the model has info on you, being told not to share it won't protect you or that data.