We can't be making public policy on the basis of PR stunts designed to help corporations effect regulatory capture. It turns out an Israeli company was behind the OpenAI and Anthropic "incidents" https://www.effort.news/irregular
The national debt stands at $39.5 trillion. Interest payments on that debt exceed war department spending. Major recessions trigger massive stimulus payments, which would balloon that debt further, which puts the entire nation into a doom loop with the bond market. So instead the economy will simply not be allowed to fall into recession. Rates will not be raised; the debt will simply be monetized.
Part of that strategy requires continual talk of rate hikes to maintain the appearance of a market.
We are clearly living through a period of national autophagy. What's happening all around us is causing a great deal of confusion, so I thought it might help to look at this from an economic perspective.
The federal bureaucracy was largely created by the New Deal, which depended on an expanding American empire. In 1935, US debt/GDP was around 39% and the world traded in physical gold, the Pound Sterling and the USD. By winning WWII, American institutional power was set in stone, both at home and abroad. For generations, Americans largely lived and died believing in Pax Americana.
Today the American empire has begun to visibly shrink. US debt/GDP is at 123%, the USD is slowly losing its reserve currency status and the prospects for American expansion are low. Decades after Reagan and Thatcher, the bond market is in total control of public spending. America is locked in an intense security and technology race with a peer competitor. Is it any wonder that institutions are being remodeled? The economics of an empire determine its laws and the shape of its institutions.
None of the above is an endorsement of empire. Creating, maintaining and unwinding them always leaves a stack of bodies too horrifying for most people to consider. Don't forget that if this stuff makes you rich.
I think it's a good thing to not bake IEEE 754 too deeply into RISC-V.
I really want to see hardware built around posits. This is not because they're necessarily superior to floats (they aren't in all use cases), but just because we need some diversity. Too much standardization is bad for innovation, and not everything that's settled should remain that way.
I have a home in Seattle, so I get to see intersections and signs that are so confusing that I can't understand what I'm supposed to do when I encounter them. On residential streets around my house, people put up all these "20 is plenty" signs, close off streets, etc., have speedbumps put in, place their cars in driveways that block off the sidewalk, put in traffic cameras in school zones, etc.
My wife's car has FSD (it's a Model X Plaid with HW4) and it does a great job at navigating all of this madness. The magical milestone for autonomy is here already, and it's a part of my daily life. I drive my girlfriend to her job and back four round trips a week, from Ballard to Green Lake.
Three thoughts come to mind:
1) At some point a federal limitation will be legally imposed on the kinds of special cases residents can force on the roads and/or we'll just shrug and let the tunnel-boring machines go under our homes.
2) I think the Boring Company tunnels would be a lot better if they had a grid topology and a story that's more than just lip service about public use.
3) It would be really cool if people got together to take over some of the surface streets and made old-timey functioning streetcars from 120+ years ago. The city and state governments could support this by simply ceding control to worker-owned cooperatives that build and maintain their own streetcars. Ballard was going to get a train that connected it to the rest of the city by 2039, and now that plan is dead because it can't be sustained for financial reasons (public planning is rife with linear assumptions and other forms of myopia).
The claim is that compilers were f(x) -> y, and LLMs are f(x) -> P(y | z1 | z2 | ... z3).
But how were various combinations of popular programming languages, operating systems and hardware platforms not effectively f(x) -> P(y | z1 | z2 | ... z3)? Suppose you were quick on the take and were writing in Unix and C in the early 80s and found yourself porting your program from a PDP-something to an 8088 PC, or to a 68k Mac, dealing with DOS extenders, printer drivers, different versions of C (remember K&R style?) or C++? Remember MFC? The evolution of the STL?
LLMs are similar to that maelstrom, just on a faster timescale.
When you're an adult and you can't afford surgery with skilled surgeons, you get shunned by the less accepting parts of society, which leads directly to increased morbidity.
When you suppress treatment before adulthood -> higher surgical costs.
When you don't treat it at all -> quiet suicide, causes get buried in the statistics.
I lived through all of these things. It was an impossible childhood. I lived with this in total secrecy and agony until my early 40s. I'm 45 now, and I've had all my surgeries. I know what hundreds of thousands of dollars buying the best care can bring. None of that is available in Finland, where this study was done.
(I had the luxury of making my peace with my father before he passed.)
The study involves a large number of people over approximately a quarter century, and its methodology is very robust. I don’t believe it’s an argument against recognition and kindness towards gender minorities, either intended or de facto.
It also focuses specifically on young cohorts so there are no people who loved in the closet for half of their lives. The range of medical transitioning described is by no means limited to surgery, and I don’t believe that therapy and hormones work differently in Finland compared to the rest of the world.
The primary takeaway here is that our understanding of gender and how to alter it is very poor and so treatment outcomes are often poor. That’s a call to more research and more efficacy, not calling trans people names and shunning them.
I understand why you’re defensive, it isn’t an unreasonable stance, but please do read the study at least.
It's telling that there isn't a single mention of surgery in the study or the interview.
Facial feminization surgery (FFS) exists but is extremely rare in Finland. After puberty, hormones alone can't change bone and cartilage structure, nor can they make vocalization congruent. Only surgery can do that.
Comparing what life was like before I had FFS+VFS, and I remember the way people treated me. After puberty locked in my bones, expecting life to be easy with hormone therapy alone simply wasn't realistic socially. Most people were incredibly judgemental. Life is a lot easier when people can't instantly spot you as something they see as abnormal.
I think you have a point, and of course no one study is going to be perfect. At the same time though one of the major strengths of this study is that it looked at large numbers of both MtF and FtM patients, and the results seem consistent, with both MtF and FtM sharing similar outcomes. If FFS was the key variable that was being ignored then you'd expect to see a stark difference between those populations.
The mean ages of participants also trended young, mostly still in their teens, which means that quite a few of the participants were receiving treatment before or during early puberty.
The point is to protect Apple's control over its APIs and business model. I'd be shocked if they're not busy figuring out how to use LLMs to shore up app review right now.
The war on general-purpose computing isn't good for human well-being. All of this stuff concentrates power in ways that are genuinely terrifying. Maybe the point is that we notice it, and understand the suffering it causes?
reply