Consider reading the thread more carefully. I'm arguing that it's foolish for people to be dismissive of parents as a political force even if they perceive the parents' disinterest in anonymity and privacy to be cranky or reactionary.
No, that is not how representation works. The representative represents, i.e they act in the absence of the citizen. They are not the human equivalent of an email broker.
> representatives are only consumers themselves
No, they are informed experts, or have informed experts at their disposal, to deeply understand the subject matter at hand so they can weigh it against the broad wishes of whom they represent. Furthermore, they have the time and responsibility to use their knowledge wisely.
> but dictators who step up and act for the people
I'm not sure why you keep bringing up dictators. Dictators don't step up and act for the people, that's what representatives do in a democracy. Are you from the US by chance? In that case I can understand why this concept is new.
> consumers love nothing more than protecting market incumbents
This is like saying: consumers sure did love dying from polio before the vaccine was invented.
The HPV vaccination eliminates most forms of cervical cancer. Australia reports a 95% drop in cervical cancer in the cohorts that got mandatory widespread HPV vaccinations. I didn't look into how that changed pap smears.
I don’t doubt your experience, but many people are intimately aware of the use of proper Unicode quote characters, in any reasonable font they choose. For me, one of the first things I learned when using LaTeX is how the quotes are transformed from the source to the typeset document; since then I’ve become extremely sensitive to the kind of quotes I see.
This analysis might benefit from a multivariate regression. You mention a few different explanations for why performance differs and if you could get solid numbers for those you could try teasing that apart.
Also a small note: the axis on one of your plots alternates between 4% and 5% increments whilst holding the ticks constant. Maybe because of rounding?
What's really interesting is that if you scale it to higher densities (eg 3nm and stacked die) with ComputeInMemory for fp8 you can reasonably start to fit 30B-70B models. With MoE and multiple stacked die, just like HBM, you could fit an open weight near frontier 1T model (like GLM5.2) at similarly much lower power <10kW and high token rates >2ktps. For running a bunch of agents where fill rate and speed/latency are important it may not matter that you're 6-18mo behind on weights. The process for the chip design could be largely automated, and new silicon pumped out as new weights are available (with a 3-6mo delay).
This is so cool but for me safety remains my priority first in place. So, if it violates my registered rules, and does what it wasn't supposed too, then what next.
Today AI agents are becoming difficult to safety most especially in critical sectors like banking, a mere mistake may lead to loss of cash or life if it's a hospital
My question is how safe is it for critical sectors
To give him credit, it's amazing that Mark Zuckerberg managed to take an easily defensible position like "we shouldn't allow AI to be owned by a couple of tech monopolies" and still manage to find so many individual points and arguments to be completely clueless on.
My personal favourite:
> Before the industrial revolution, 90% of people were farmers growing food to survive. Advances in technology steadily freed much of humanity to focus less on subsistence and more on the pursuits we choose.
Over a long enough time span, the industrial revolution has certainly improved living standards but the industrial revolution itself was l, for many, an absolutely horrific time to live. Life expectancies in cities plumetted and we saw the birth of workhouses.
It's such an odd example to choose for trying to suggest "technology x will help everyone in a non-complicated way". He might as well have picked the atomic bomb.
there's probably a box to check on the "be attractive to investors" form and then you can also put the ethics department in HR docs as an escalation path to make emo employees feel better. I bet the ethics people get told how important they are when hired and the team is stood up but then get upset when they're continually bypassed and quit. I bet the bait/switch hurts a lot.
>We let them in, so they can work our fields and process our meat and whatever else. We shouldn't be surprised they want to buy cars, own business, and do other completely normal things.
So if the city/state/federal government lets a drug addict/dealer population to exist for too long (despite their best efforts at cracking down), they'll eventually get grandfathered in, and we can't enforce anti-money laundering regulations or dismantle their crack dens?
Thanks for warning about the dangers of hiking or having sex, but just because people can die by multiple reasons does not make this toxin one inch safer.
If the pros are discussed, the cons should be discussed also.
The number of people that has sexual activities repeatedly for three decades without dying is still much bigger than the number of people that licks toads repeatedly for three years without dying. One activity has been much more tested than the other.
I don't know how people can say this with a straight face. Nvidia was selling desktop-grade ARM SOCs before Apple Silicon was ever announced, specifically for edge robotics, computer vision and ML.
The absolute fastest desktop Mac GPUs cannot beat an Nvidia laptop GPU in prefill or inference speeds. Apple Silicon is a non-entity for professional datacenter deployment and arguably unusable for frontier models at agentic context sizes. AMD is Nvidia's primary worry, and they're not doing much better in terms of GPGPU SOC compute.
Indeed. At uni, starting in my second year, I worked as a programmer (no remote work, office at 1hr commute from uni dorm) on top of my studies. It was hard, but doable. My major was applied math, so the CS part was the easier bit, but the math, especially Analysis, Numerical Methods, Linear Algebra, and Analytic Geometry they kicked my butt. We also had a really bad prof for the Math Logic class and no textbook, only lecture notes and his lectures were unstructured shit :(
The biggest problem wasn't the workload itself, but how inefficient learning was back then. It was the early days of the Internet, no coursera, no Google indexed uni sites, no LLMs, no online tutors. I only had a mediocre textbook and some notes I borrowed from friends when I couldn't attend lectures or labs. It took ages. I’d burn hours trying to parse two paragraphs in the textbook, only to ask mates who didn't understand it either.
Now, we have an incredible amount of information & tech at our fingertips: LLMs, Google NotebookLM, online tutors, dozens of textbooks, and sites like Anna's Archive packed with high-quality "free" material. With all this technology, the time it takes to understand, prepare, and master a topic is a fraction of what it used to be. It still takes work and discipline, but man... you can feed textbooks into an LLM and use targeted prompts to debug your exact knowledge gaps on the practice test you get from the LLM.
So I reckon it's easier now than before, simply because learning can be done so much more efficiently. After all these years, I still remember the hardest classes and the most stringent professors with great affection - even though I hated those bastards back then. My Group Theory prof once invalidated a definition I wrote because of a single missing word. At the time, I thought it was an injustice, but later I realized that word was crucial.
University is where young people must face hard work and reality, because it teaches them what true meritocracy is. It should be hard, it's the same way a gym workout should be hard or it’s ineffective.
I am against marijuana but I think it's clear legalizing it successfully starves the cartel of a lot of drug financing (and therefore we should both legalize it and have public campaigns against using it, ala alcohol).
Even the weird state based legalization has caused the cartels to pivot into things like avocado farming and mining.
Drug users don't seem to be overly preoccupied with what drug they use. Probably shifting them towards lower risk choices (pot, ketamine) and aggressively crack down on dealers is a winning combo.
(A lot of low level dealers are just graduated drug users but the line between buying and selling is a large one and dealers do immense harm while also shielding more dangerous criminals. Incentives against dealing incl aggressive enforcement are powerful at dissuading low level dealers, which in turn starves the criminal organizations.)
Perhaps worse than what the US had before the fall, but it's still an open question of whether Europe will hold it together or fall to the same corpo-authoritarian + revanchist-destructionist movements that are killing the US.
