Security is often an excuse to block other teams to do legitimate work and so often it's fairly braindead. Security IMO needs to get it's act together, passkeys is a great example of security gone wrong from a UX design perspective because you can't hold them to the same standards as product or infra teams, they have the special privilege of breaking things and it increasing their metrics.
Tell them to make a better UX and they lose their minds in a huffy puff of fake crisis mode or get avoidant with stonewalling 'secret security stuff' that you can't hold them to account for. Or eat 50% of developer machine performance for "endpoint security" and the carnival of sadness goes on and on.
Signal is an example of security as a product that was actually designed for user UX in mind to give one example.
Yes we do, continue keeping it a faux pas to reduce the over verbose LLM speak put elsewhere and ask people to just share the original prompt and save us all time. Label your LLM usage to respect other people's time.
If you've used UIKit to any large degree, how bad SwiftUI is, almost a decade later, continuously punches you in the face. About 20% more time to write initially, for 90% less bugs and potholes, and I've tested this with iOS devs who never used UIKit before too. Now with AI, there is even less of an excuse IMO.
Leave SwiftUI to the settings pages. The gulf between AppKit and SwiftUI in macos desktop I'm not as sure about.
I have an M1 Pro, and a M4 & M5 max to play with at work and the speed difference is very significant between all 3 machines, the M1 Pro is far slower, and the M5 is significantly faster than the M4. And a windows 3090 beats all of them but eats twice the amount of power per token. This is all running the same 24GB memory friendly model with LM studio.
I played with classifying and summarizing my entire email history (per email) with small models, but that only took about 12h of GPU time at most. Using a coding agent cli wrapper in that case is far slower because of all the spin up cost and the system prompt they inject even if you want to turn it all off.
If I used an actual direct API it probably would've been much faster, but I'm doing it for hobby / fun reasons. You also get to fiddle with a lot more params.
Anthropic is the AI doomer / safetyism lab, and Hinton is one of the patron saints of 'rationalist' AI doomerism.
AI doomerism is psychologically attractive to "people with autistic cognitive traits, including dichotomous (black-and-white) thinking, intolerance of uncertainty, and a tendency toward catastrophizing". They are pascal's mugging themselves, to ironically use one of their terms. It's fundamentally a cognitive distortion.
"What if AI doom is all fear-mongering, and we create AI less prone to make up dangerous stuff or mistake buggy goals for real ones" (which is what alignment is) "for nothing?"
Even if Yudkowsky is autistic, you're muddling the condition. Autistic people have a *practical* intolerance of uncertainty in the moment (everything unexpected from a noise to a missed turn can be a jump-scare in their day-to-day activities), but often they're absolutely fine with intellectual uncertainty, unconventional ideas, abstract ambiguity, nonconformity, etc. Indeed, one of Yudkowsky's main things is Bayesianism, i.e. being precise about uncertainty.
Yudkowsky's reported P(doom) is somewhere around 90%, which is very much in the realm of "we might eventually be able to figure this out, *but we're not even close to ready so for the love of everything slow down so we can figure this all out*"; the book title comes from a long tradition of authors noticing you need to beat readers over the head with your point for them to notice it.
Anthropic (like at least also OpenAI), appears to think they can solve the problems that Yudkowsky has found. They're a lot more optimistic than him, but they take these problems seriously.
For his work on AI, Hinton got a Nobel prize in Physics, a Turing Award, the inaugural Rumelhart Prize, a Princess of Asturias Award, a VinFuture Prize, and a Queen Elizabeth Prize for Engineering. Calling him a "patron saint" of "doomerism" is like calling Paul Krugman (Nobel laureate in Economics) a patron saint of "Trump Derangement Syndrome" on the basis of what he says in his YouTube channel: a smart person's considered opinions are worth listening to even if you have not got time for the details, because you can be sure someone else has considered the details and will absolutely be responding to even an i missing a dot.
The problem is that effort spent to reduce the "risk" of creating an evil god who tortures us all for the rest of time doesn't actually produce outcomes that reduces the risk of things like widespread job loss or hyperaggregation of influence and money.
"Oh we'll at least get some side benefit" is not actually what is coming out of the endlessly circular forums talking about the apocalypse.
Even if there was no overlap*, that would be like criticising the green movement for not focussing on working hours and pay like trade unions do.
Different people can care about different things; it's good that each of us gets to focus on what motivates us, rather than all chasing the same thing, because when multiple teams do all chase the same thing typically only the best few of them actually make a difference.
* as it happens, there is some overlap. Knowing more about how a narrow utility function behaves outside distribution is useful for both capabilities and safety. We're not even at the stage of being able to make AI not kill random subsets of the users with bad advice, nor reliably prevent users from falling into delusions of grandeur, let alone giving AI a reliable sense of liberty and the pursuit of happiness to maintain.
Much like a lot of LLM usage burns tokens so that mediocre people can hallucinate that they're doing something brilliant, Yudkowskyism is just a lot of empty verbiage for the purpose of building a sex cult around a plump gnome. Reusing his nonsensical and poorly defined terms but failing to get the benefit of the sex cult really misses the point of the entire exercise.
I have a younger family member who is AuADHD and is under AI & other doomer delusions. It's effected her life to a large negative degree at this point. It doesn't matter if Yud is autism leaning, as a movement it's far beyond him. AI doomerism and other doomerisms is Q-anon for smart, moral, austism-leaning people at this point, which is incredibly sad. Because it is something that really affects smart people mostly, you can't use the qualifications and intelligence of it's proponents as measures of it's validity, you need to look at the thought distortions directly itself. It consumes lives and it definitely can hurt people.
The people who've made the biggest contribution to creating a better world over the last 50 years have been the Chinese; powered largely by coal and petroleum. And in one of the most ironic results in the 21st century, they're now the leaders in solar panel production on the back of the largest investment in fossil fuel energy in global history.
The comic ran into the same problem as the climate change movement in general - they proposed ideas that generally made people worse off. And if measured in terms of CO2 emissions achieved nothing except pushing wealth creation to Asia. Which, in fairness, is probably appreciated by the Asians.
BYD had release the first plug in hybrid the year before.
The Beijing Olynpics had made air pollution a hot topic in China in 2007-8.
Wind power had accelerated after their 2005 Renewable Energy Law.
Solar panel production rose around this time, taking over the market from European manufacturers when the Financial Crisis hit and they pulled back investments.
So China at that time, was doing all the things on the cartoon's presentation list, and has benefitted greatly from them.
Many people in Europe want to see green energy transition. But no transition is happening in China.
"
“We see addition, not transition,” said Yasheng Huang, a professor of global economics and management at the MIT Sloan School. “China is building alternative sources of energy as well as fossil energy sources, simultaneously. In terms of the global footprint on CO2, China is emitting twice as much as Europe and the United States. I don’t think there’s a transition going on.”
"
What an embarrassingly ill-informed thing to say. But when the guy wrote a book in 2023 about the fall of China, he kind of has to say that doesn't he, even as he lives through the fall of the USA.
He's called out in the sub-head as an "expert" but what is he an expert in? Renewables? Energy policy? No, he's an expert in saying that China is too state-led. Why would an expert in that want to downplay their success, apart from all the obvious reasons?
"
For Beijing to achieve those goals, Climate Action Tracker says China needs "clear targets for coal consumption reduction" in its new 5YP. However, the economic roadmap released in March was not "explicit about how fossil fuels will be constrained," said China analyst Qi Qin of the Finland-based Center for Research on Energy and Clean Air.
Though Chinese President Xi Jinping promised in 2021 to detail a reduction in coal energy use in the 2026-31 plan, it contains "no clear phase-down plan, no clear fossil fuel cap," said Qin. "The language is much more conservative than many people expected," she told DW. One reason is the continued influence of the powerful coal lobby on Chinese government policy.
"
> New Chinese government guidelines on fossil fuels released on April 22 support the view that the country is willing to move away from finite fossil fuels, strengthen energy independence and still achieve its climate targets, says Qin.
> "The new central guideline talks about strictly controlling fossil fuel consumption, reducing coal and controlling oil. It still leaves room for flexibility, but these are concrete policy levers," Qin said of the document, which also indicated a desire to increase clean energy consumption.
Elsewhere Climate Action Tracker on the USA:
> The Trump Administration is pursuing an executive and legislative agenda to systematically repeal targets, policies, and funding for climate change mitigation and science. The administration is actively obstructing the buildout of renewable energy, while encouraging the production and consumption of fossil fuels, completely reversing the Biden Administration’s course on climate action. This is the most aggressive, comprehensive, and consequential climate policy rollback that the Climate Action Tracker has ever analysed.
All of which, even the bits quoted to claim "no transition is happening", support my original contention that all the things mentioned in the cartoon were being strongly pushed by China in 2009. They have only gained momentum since and they've profited from doing so.
I agree that the Chinese 5Y plan is better than the US policy "Drill, baby, drill!". But how much exactly, we will see.
Lets drill more into details:
"
Against this backdrop, coal has re-emerged as a critical stabilizing force in China’s power system. This helps explain the relatively cautious policy signals embedded in China’s 15th Five-Year Plan. While the plan reiterates long-term decarbonization commitments — including reducing carbon intensity by 17 percent and raising the share of non-fossil energy to 25 percent by 2030 — it still stops short of setting explicit timelines for coal or oil consumption to peak. This reflects a deliberate effort to preserve flexibility as Chinese policymakers balance energy transition goals with near-term electric system stability.
In practice, China’s coal production has rebounded significantly. After nearly a decade of supply-side reforms that kept output around 4 billion tonnes annually, coal production rose sharply following the 2022 power shortages and has continued to increase, reaching a record 4.85 billion tonnes in 2025.
"
I would like to note "reducing carbon intensity", doesn't mean reducing total carbon emissions, it's reducting carbon emissions per unit GDP.
This is in strong contrast to EU
"Specifically, the EU has a legally binding headline emission reduction target of 90% by 2040 relative to 1990, with a domestic target of 85% and up to 5% of international carbon credits."
Something that has been largely forgotten about is that it used to be routine to see pictures of smoggy Chinese and Asian cities, this was a problem for them that they solved. I can't help thinking we can't get this kind of preventative action on any large scale, we need to have severe issues first and that's not accounting for longer term/cumulative effects.
"Over the past years, the government has implemented various methods to improve the air quality in Northern China. Sandstorms, which were quite common 15 years ago, are now rarely seen in Beijing’s spring thanks to afforestation projects on China’s northern borders. The license-plate lottery system was introduced in Beijing to restrict the growth of private vehicles. Large trucks were not allowed to enter certain areas in Beijing. Above all, the coal consumption in Beijing has been restricted by shutting down industrial sites and improving heating systems. Beijing’s efforts to improve air quality has also been highly praised by the UN as a successful model for other cities. However, there is also criticism pointing out that the improvement of Beijing’s air quality is based on the sacrifice of surrounding provinces (including Hebei), as many factories were moved from Beijing to other regions."
The US had the same issue and fixed it through federal and state environmental regulation. It just happened in the US 100 years before it happened in china Heavy pollution is what lead to the environmental movement that started back in the 60s and that led to the creation of the EPA and whole slate of state and federal regulation that dramatically improved air/water quality in the US. It was a slow process that took a ton of work to build a movement of support, but it can be done.
We can actually address problems when we want to. It's just pretty slow and requires people to actually give a shit and put in the effort to build support.
The unfortunate comparable here is that all the people who care about making sure their AI is safe, regardless of what they mean by that, are beaten to the market by the people who don't.
Just like how you significantly increased the difficulty of exams in "open book" exams in the past where the only way to pass the open book exam was to know the material well, you similarly need to increase the difficulty of other work where it won't matter if you have an LLM, because you won't pass without knowing your shit either!
The problem is that only works at the advanced courses. However people need to learn the basics before they reach that level, specially when they are starting and are in many regards below the LLM's baseline.
Need to type? Computer labs (“test taking labs” idk) are back baby. Simple machines, no Internet.
Pretty sure that solves 90% of the testing problem. If somebody is overly reliant on LLM’s and refuses to learn, they’ll pay with their grades on the big assignments. Bummer for teachers who don’t love blue books, but I’m sure it’s a hell of a lot better than trying to sniff out LLMs and constantly mistrusting your students.
Tell them to make a better UX and they lose their minds in a huffy puff of fake crisis mode or get avoidant with stonewalling 'secret security stuff' that you can't hold them to account for. Or eat 50% of developer machine performance for "endpoint security" and the carnival of sadness goes on and on.
Signal is an example of security as a product that was actually designed for user UX in mind to give one example.
reply