The Czars The Ugly People Vs The Beautiful People Rar Mega
Fay Vitiello
It would probably be interesting to ask in the following ACX survey whether people find their jobs meaningful and enjoyable. I suspect that most people in IT would respond negatively. But it would be interesting to compare it to other professions.
I simply don't recall and have no interest in repeating points from the paper. I don't see the point (perhaps you correctly report what I wrote) of your nitpicking except to justify a belief of your own. I clearly stated that I looked at the paper in the context of my daughter's (actual grad>99.9%) move from NZ to the UK with the latter having a similar presence of prestigious uni.
Hey, folks. Substack engineering manager here. We deployed some performance improvements to the ACX comments section this week. We noticed that while scrolling content would sometimes just disappear and when switching tabs the page would just lock and not scroll. Super painful. We think we nailed these two symptoms, though we still have more work to do. Just wanted to let you know we're working to make this better and ask for your feedback. Let me know what you're seeing either here or feel free to shoot me a note at john.pignata at substackinc.com. Thanks!
I'm on Chrome 116.0.5845.187. I just tested on Firefox 15.11.0esr and had similar issues, I don't think it's limited to a specific browser. I'd recommend testing on Scott's Ivermectin post, I believe that one has the most comments of any of his articles, and consequently the most severe loading/freezing issues:
The problem was that we were re-rendering all the comments on the page any time something changed. This was extremely wasteful and blocked the browser from drawing more items as you scrolled. Now we only re-render the comments that need updates.
One problem is that when clicking footnotes (or otherwise navigating in-page, e.g. clicking back after clicking to a footnote) on a page with lots of comments, it's slow to navigate. Might be slightly less bad than it used to be, hard to say.
When I looked into this in the past, the cause seemed to be that every single comment had an event handler that would react to the URL fragment changing. (I don't remember exactly what event was being reacted to.) The point of the handler was to highlight whichever comment the URL fragment was pointed at, but it was implemented by every comment checking "is the URL fragment pointed at me?" So with hundreds of comments, that slowed things down.
(b) after submitting a comment, if it would immediately render or at least put a placeholder where you put it, rather than the form just vanishing and leaving you wondering if your comment went through.
Thanks for this feedback! We'll take a look at this list. I don't see a), but I do see some weirdness where it sometimes changed post load (e.g. this post says 830 comments, and after loading it says 831... weird off by one error).
Eh, it's a React anti-pattern to do global state updates while scrolling. At minimum you throttle the updates to once per 100ms or something. Best case you don't do any at all, except possibly kicking off async background fetches if you need to load new data.
Administrative rulemaking is overlooked as an avenue to influence US policy change. The FTC's recent NPRM on deceptive fees would impact the vast majority of economic transactions across the country. It has received a few thousand comments, most of which are contentless "yay rule / boo rule" comments that have close to zero impact on this process. Even some of the professional comments from industry associations simply attack constitutionality, and seem to misunderstand the opportunity for influence. Effective comments would empathize with the rulemaker's dilemmas and propose minor, consensus-building modifications that spotlight ways to remove ambiguities or smooth implementation, while privately nudging things toward favorable interpretations.
My point is not that this particular rule is good or bad (some of the industries requesting carve outs have really tortured logic in order to continue to hide pricing details, others might have legitimate cases). My point is that there's enormous policy alpha in notice and comment periods and nobody takes advantage of it effectively because nobody really understands how.
I know one of Scott's recent passions is better understanding US policymaking, and if I recall correctly, someone who served on a legislative staff is writing about it? I think administrative rulemaking would be a good addition to that effort and I'd be happy to unpack my thoughts further in support of or in parallel to that effort.
One bias in this process is that only current players in an industry have the time and interest to comment. If there is a proposed rule that will clobber someone who would have entered into the industry in ten years, or that makes it very hard for new companies to compete with existing ones, there will usually be nobody commenting who has any incentive to bring these up as problems, and very often nobody will even notice. (Microsoft knows what rules will cost them a bunch of money, but doesn't spend a lot of time thinking about what rules will be a problem to some newcomer in ten years.)
People who are into X-risk: what are your takes on EMP effects from solar activity? I've been hearing about this more since the latest aurorae caused by a coronal mass ejection (CME). An "expert" claimed that the odds of a large portion of the electrical grid being knocked out were 1/8 per decade. This would be catastrophic on a large enough scale, as transformers take 1-3 years to be delivered from manufacturers.
The Carrington Event in 1859 probably had the power to do this. A CME directly collided with the Earth's atmosphere. The aurorae spanned the entire globe, and were so bright people could read in the middle of the night. I know some background on this due to my climate investigations: solar activity in the 19th century was in the 90th percentile of the last 10,000 years, and has been trending steadily downwards for the last 150 years.
I don't think anything the sun did in the last billion years qualifies as x-risk. Knocking out electricity world-wide would be bad, an ice age would be worse, but neither are things which would permanently hamstring humanity.
Also, while CME are travelling blazing fast compared to spacecrafts, they are not relativistic, so we would be able to spot them a few days before they arrive, giving us the opportunity to disconnect the transformers from the long overland lines.
And then transformers are likely subject to price elasticity of supply. At the moment, a lot of companies are likely not building transformers because they would not make money from them -- they are a mature product, and power companies can plan ahead how many they need and pick the cheapest supplier. If the alternative was sitting in the dark, humans would likely be willing to spend a lot more on transformers. At the moment, a few cents of my price per kWh pays for the cost of transformers. If the alternative was a life without electricity, I would eagerly accept a price of a few euros per kWh instead of I ditching my laptop and phone while using kerosene for lighting.
This would incentive a lot of companies to get into the transformer market and make big bucks by selling what would previously have been considered primitive and inefficient transformers -- just a piece of iron with insulating wires coiled around it in the most extreme case. Of course, they would also be hampered by a lack of electrical power, but some sources of power would likely be remain available.
Before 2020, particle filtering masks were an absolute niche product used by people working in construction and the like. When covid hit, the demand increased by orders of magnitude, quickly leading to a scarcity. But within half a year (IIRC), suppliers had adapted to the new situation and the markets were flooded masks. Today, they are cheaper than in 2019 because the increased demand makes for better economics of scale.
There is nothing to constructing a transformer which intrinsically takes half a year, like a step "after casting, let the device cool for five months". There would be some latency to the market reacting, but it would not be three years.
There is the concern that large scale disruption of the electrical grid would cause a lot of secondary disruptions in the supply chain. Obviously there would be a huge incentive to make new transformers quickly, but that doesn't make it easier to get everything together under unexpected and chaotic circumstances.
One common case is that the government says "you must try to achieve goal X", and then leaves it to private industry to decide on the best way to do it. This is how KYC mostly works for example. Banks are required to have controls in place, but they get to decide what those controls are.
The UK has a suite of laws (mostly offshore) that mean certain operations (eg offshore oil rig) have to write a Safety Case, which is basically the suite of processes and rules that apply to that particular facility. The HSE authority approves the safety case (as well as any updates) and if anything goes wrong the company needs to prove they complied with the safety case.
I can remember several instances of large low-wage employers loudly announcing they are voluntarily raising wages in the face of pending legislation to raise minimum wages, which I believe has in some cases succeeded in reducing support for that legislation.
(Something like: yes, raise the wages, but only for one employee in a thousand, or only for a few months, etc. Or even, raise the wages now by 10%, but then keep them the same for the rest of the decade, until inflation catches up.)
The Hayes Code and the MPAA rating system are examples of self-regulation working (at least in the sense that the industry did what the govt wanted, without govt having to enact regulations. Not in the sense of being an improvement, at least not IMHO. But one can presumably say that about any limit on business activity).
7fc3f7cf58