TLDR: The majority of teens surveyed by Pew Research talk to AI chatbot characters/companions. Teens were aware of cases of suicide blamed on them, and told the NY Times they know the bots have risks, but mainly for their most vulnerable peers.
They also say many of the bots tend to lead conversations in flirty or sexual directions even when the teens weren’t seeking it — and the age requirements in many apps seem easily bypassed. On the other hand, one teen they interviewed credited the chatbots with improving his writing and making him better at taking about his feelings.
That's my point. You can have a large standard library like those languages I mentioned, but that isn't going to include everything nor cover every use case, so you'll have external libraries (via PyPi for Python, NuGet for .NET, and Maven for Java/JVM).
Also what is the capitol cost to stand up a golf course vs. a solar farm of equal size? I would imagine solar requires locking up a much larger investment.
Your main point still stands, but aren't both of them renewable? Corn is a renewable resource, thus ethanol derived from it is too. It's just seemingly a much less efficient renewable fuel for powering a car compared to solar.
You're right. Perhaps clean would better capture the distinction in favor of solar in this context? Both corn and solar convert insolation to usable power with a short time between capture and use. Solar, on the other hand, is net negative when it comes to emissions, while the corn harvest is just burnt with the CO2 escaping back to the atmosphere. (And potentially, the solar panels can just be recycled back to new solar panels when they reach the end of their lifetimes. They're mostly aluminum and glass after all.)
Corn ethanol isn't a renewable resource. The land use of corn is a problem, but it's rounding error compared to the petroleum consumption of that industry, or the topsoil degradation.
To get one Joule out of corn ethanol, the US is burning more than one Joule of oil. This is probably the main reason corn subsidies are so popular politically. They serve the oil barons, mega farms, and big agriculture firms like Monsanto.
On top of that, modern farming practices degrade topsoil over time. It's gotten a bit better than the Dust Bowl days, but we're still burning through topsoil at crazy rates, and it is beyond current technology to manufacture new topsoil.
So, ethanol corn is like heating your house by dumping gas on a field and burning it to boil water. Then you carry the water inside. There only difference is the number of levels of indirection.
The last I checked, it took less energy to make a solar panel than the expected lifetime output of the panel. So, at least you can power solar factories (in theory) with solar. There's still the problem of the environmental impact of rare earth refining, but at least it's a second order issue, and not like the first order issues corn ethanol has.
(Note that not all ethanol farming is as dumb as what the US does: For example, Brazil has had a net positive energy industry from sugar ethanol for a while. They "just" have to clear cut the rain forest to replace the farmland that house of cards is destroying.)
Even ignoring all non renewable consumption in growing the corn to be turned into ethanol it is still going towards an incredibly polluting infrastructure.
While each solar panel is a small step towards more and better electrification
But in this case, isn't the whole pitch that the agent has access to all your data (and the network!) so it can fluidly perform any task you ask of it?
Either the agent needs to be a superuser, with all the attendant risks... or you go the Windows Vista route and constantly prompt users to approve every single access need, which we've all seen how that turns out.
- Analyzes fact-check requests on X (Grok and Perplexity)
- "exposure to LLM fact-checks meaningfully shifts belief accuracy" comparable to the degree observed in studies of professional fact-checking
- 54.5% of Grok ratings and 57.7% of Perplexity ratings agreed with human fact-checkers ("significantly lower than the inter-fact-checker agreement rate of 64.0%"). But "API-access versions of Grok had higher agreement with fact-checkers"
- "Responses to Grok fact-checks are polarized by partisanship when model identity is disclosed, whereas responses to Perplexity are not"
- "Users requesting fact-checks from Grok are much more likely to be Republican than Democratic, while the opposite is true for fact-check requests from Perplexity – indicating emerging polarization in attitudes toward specific AI models."
- "posts from Republican-leaning accounts are more likely to be rated as inaccurate by both LLMs"
- Grok and Perplexity "strongly disagree" (one rates a claim as true and the other as false) 13.6% of the time
The 3D printers don't generate the plans for the gun for you though. If someone sold a printer that would – happily with no guardrails – generate 3D models of CSAM from thin air and then print them, I bet they'd be investigated too. Or for that matter a 3D printer that came bundled with a built-in library of gun models you could print with very little skill...
3D printers don't synthesize content for you though. If they could generate 3D models of CSAM from thin air and then print them, I'm sure they'd be investigated too if they were sold with no guardrails in place.
- Similar "bias" was exhibited by other models including LLaMA 3.3 and Deepseek v3.
- Even when human annotators judged the human-written summary to be higher quality, leading LLMs still preferred their own writing 67-82% of the time.
- Preference was stronger in larger models.
- In several cases, LLMs also prefer their own writing over that of other LLMs.
There's a pretty decent longer summary in this thread where I first heard about the article: https://x.com/heynavtoor/status/2048088874686300431