It says "more like", not "like". Javascript now performs more like a statically-typed language, as one example. That wasn't always the case. It used to be painfully slow — and was so when Go was created. The chasm between them has shrunk dramatically. A fast dynamically-typed language was a novel curiosity when Go was conceived. Which is why Go ended up with a limited type system instead of being truly dynamically-typed.
A committee from the beginning would definitely prevent something from really ever starting. Could you imagine Linus working under a committee to get Linux running?
At some point, you do have people that need to step back. If you turn it over to another single person, they could pivot and "ruin" the product. By turning it over to a committee, hopefully, any ruinous ideas get overruled. At least in theory
The other issue is that it's expensive. You can put a video on youtube for free and they carry the cost and cover it with advertising. If you self-host and your videos get a LOT of traffic it gets expensive quickly.
I feel people are often unreasonably scared by this sort of thing, so here are some numbers to give perspective and aid in decision-making.
Bunny CDN charge $5/TB for their volume network, which should be pretty good for video distribution, reducing after 500TB/month.
At a bitrate of 5Mbps (respectable for 1080p, significant overkill for more static types of content, as technical stuff will tend to be), 1TB is 444 hours. If, like OP, you publish 90-second videos, that’s 17,777 complete watches per terabyte. Depending on your situation, that might sound like not much or like a lot.
Put the other way round, at 5Mbps and $5/TB, each watch-hour costs $0.01125, a bit over one cent, and it takes 3,555 people watching your 90 second video to cost one cent.
For the sort of scale that most people are dealing with, it’s simply not an issue.
I don’t know if bots upset this balance. They may.
If you actually are spending more than a terabyte per month on it, then for technical audiences at least, I suspect that if you invited donations to specifically cover hosting costs (something along the lines of “I host these videos myself because ads and relying on YouTube are both bad for society; if you feel inclined, you can donate to help cover the cost, currently about $X/month”) you’d very quickly get a surplus. Or for longer-form content, charge something for 4K video (which costs 4.5¢ per watch-hour at 20Mbps and $5/TB) and let that subsidise the free 1080p (costing 1.125¢ per watch-hour) stream.
(On the $5/TB figure: my $5/month Vultr VPS includes 1TB per month, and charges $10/TB after that. Some VPS providers include a lot more; a Hetzner €3.49/month VPS in Europe includes 20TB then charges €1/TB. But remember, if you host video from one point only, that it is unlikely to work well for people halfway round the planet. See another of my comments in this thread for description.)
As for storage: each 90-second 5Mbps video is 56.25MB, and at a rate of $0.01/GB/month, each one will cost you $0.00675 per year to keep. Were you to post one 90-second video every single day and keep them all online, your monthly bill would grow by about $0.20 each year.
Another way to estimate upper/lower bounds for hosting videos is by considering the grey market adult video industry. It seems like there are hundreds, if not thousands of websites providing access to video, and from a unit economics perspective I remember reading those advertising CPMs are on the order of $0.01-$0.05 even for the biggest and least illegal websites like the Mindgeek properties. So I would assume the more shadowy websites operate at a budget of less than $0.01 per thousand views in revenue.
I’m assuming their CDNs are just specialised low cost hosting providers as opposed to p2p IoT botnets, but you never know.
I wish someone had the time and motivation to do an investigatory technical deep dive into the infrastructure they use.
LOT of traffic is a big if. A lot of videos only have 1k+ views and I will believe those are mostly drive-by viewing. Using a CDN can help you if you go viral.
He says that as if he's certain it can't possibly, even a remote possibility, lead to societal collapse. First, there is no way he can be certain about that. Second, what is an acceptable probability for an existential threat? That's the real question to answer, and he didn't attempt to answer it.
Climate change could do a lot of damage it’s just not extinction level damage. Even large scale nuclear war based on current stockpiles isn’t going to result in extinction.
There’s levels of societal collapse, mass migration can destroy the existing social fabric without necessarily being that terrible. Fertility rates being so low means developed countries will likely want large numbers of immigrants.
At the other end stopping all CO2 production tomorrow would result in severe consequences. We can’t transport food to cities without burning fossil fuels. Obviously that doesn’t mean every current use case is worthwhile, but we can’t ignore the short term here.
The good news is we’re actually making a lot of progress on climate change. The electric grid being ~90% very low carbon emissions by 2050 is a realistic goal and would avoid the worst predictions.
That graph is well below earlier forecasts. Accurate predictions require more than simple extrapolations which 30 years ago suggested exponential growth. Instead CO2 per capita and especially in terms of GDP resulted in a different story.
Poor countries are rapidly becoming wealthier which is obviously a good thing. Meanwhile wealthy countries are becoming a lot more efficient with their carbon emissions. Where those lines intersect is what matters in the near term because poor countries aren’t copying 1950’s technology. Skipping power hungry CRT screens and inefficient engines etc just makes economic sense. China emissions spiked as they industrialized but they are currently minimizing their investment in outdated fossil fuel based technologies in favor of solar and EV’s etc.
Co2 levels are the best metric for progress, and progress would be a steady lowering of the rate of increase. We are not seeing that, and AI's power hunger is likely to make it worse despite all of the positive things you mentioned.
CO2 levels tell you nothing about future progress, the legacy of soon to be removed infrastructure built 40+ years ago still impacts it. Meanwhile the vast majority of power brought online in the last decade is ultra low carbon. We know how that story plays out.
AI powered by renewables has ~zero impact on the climate.
I hope you're right, but I'm not optimistic. The demands for power by AI is immediate and extremely large. We can't bring nuclear online fast enough so it will likely keep co2 output high.
Assuming you are right, when would you expect the rate of atmospheric co2 increases to start to decline?
I expect the global peak annual CO2 emissions from fossil fuels and industry to occur sometime in the next 5 years largely dependent on economic activity. I’d give it something like 90% odds.
CO2 from land use (deforestation etc) has already dropped by half since the 1960’s, but I know less about that so I’m unsure of the details. That said it’s well under 10% of total emissions so probably not a major factor for now. https://ourworldindata.org/grapher/co2-land-use?mapSelect=~C...
Is infinite growth really a positive thing on a finite planet? Are native populations replaced physically and culurally the answer to making an imaginary number go up annually?
My point was about a declining global population concentrating in a few areas not being a major issue.
Globally more babies were born in 1985 than 2025. Population growth at this point is all down to people living longer but that’s a one time correction, we’re already in a steady state situation and heading to decline. https://ourworldindata.org/grapher/number-of-births-per-year
Non-zero describes the chance of everything, sure. Infinite improbability drive and all that. But the chance of climate collapse causing social collapse is pretty much just a function of how bad we let it get measured in degrees C.
Your example is infinitesimally small, climate change is not. But you didn't answer the question either, what probability of societal collapse do you think is acceptable?
Having grown up in the 60s and 70s, I'd say people took nuclear war seriously. People had different opinions on how likely it was and whether it was an extinction event, but there was near unanimity that it was "a really bad idea." The obvious difference was that was impossible to doubt it was man-made and it wasn't something that slowly built up over decades--there was no way to say it was "normal"
I think it’s a fair point to say it was considered as a serious threat to certain countries(US/Russia/UK/China etc). And militaries certainly did prepare for it. But other than Switzerland with all their bunkers - which society in all facets prepared for it, really?
WW2 with it’s restrictions & rationing, and almost all civilian economy/effort being redirected to the military is I think what a lot of people are wanting in my honest opinion.
People say this all the time, but I've never seen any data proving it's true. Should be rather easy too, I'm at a big company and different teams use different languages. The strictly typed languages do to have fewer defects, and those teams don't ship features any faster than the teams using loosely typed languages.
What I've experienced is that other factors make the biggest difference. Teams that write good tests, have good testing environments, good code review processes, good automation, etc tend to have fewer defects and higher velocity. Choice of programming language makes little to no difference.
For starters we have cancelling and ripping up clean energy projects, introducing serious uncertainty about tarrifs, increasing unease for any immigrant (and even non-white citizens).
Avg monthly "clean energy" jobs so far:
- 2023: ~3.38M
- 2024: ~3.45M
- 2025: ~3.42M
Any impact to the economy due to "green job project cancellations" wouldn't have surfaced yet.
Uncertainty about tariffs: it was already on the books to earn $2.3T (conventional) / $1.5T (dynamic) over the next 10yrs, and inflation is far more under control than it was last year despite the tariffs. Of course legal challenges have disrupted it now. This is also completely ignoring the massive investment foreign countries/companies are making on US Soil to employ US employees.
Edit: having said that, "these economic problems are ENTIRELY the result of bad policies". That's a very bold claim, which warrants some strong evidence. No one has provided any strong evidence that it's ENTIRELY from bad policies.
I don't the current economy is "entirely" the result of bad policy, but I do think it is the primary contributor. I read/listen to a lot of economic commentators as an interest of mine, and there is pretty broad agreement that tariffs are the main cause of inflation failing to tick down to target this year. There also seems to be consensus that the manufacturing sector has been harmed by tariffs and immigration policy rather than helped.
In my opinion, the AI hype cycle has temporarily buoyed the economy from more serious pain. If significant economic gains aren't realized from it soon I think we'll begin to see that pull back.
I do, however, think a return to ZIRP by the Fed would result in a significant economic boost. Psychologically, everyone remembers how advantageous low interest rates are and I think it could result in real investor/borrower optimism, temporarily, if we go back to that. Unfortunately, that would likely mainly stimulate the demand side of the economy, and not as much supply. I don't have high hopes for how that would affect inflation.
Yes, no major disagreement here. You can also point to the Us debt/deficit as another major storm cloud on the horizon. Trump is the worst offender on that, but far from the only one.
I'm as anti-trump as anyone. He's doing tremendous harm to our economy and our society. Your anger is making you see things that aren't there. All I did was question the assumption that our economic problems are ENTIRELY the result of bad policy. Some how you jumped to some very wrong conclusions based on that tiny shred of information which wasn't even about Trump.
You coyly suggested that there were other contributors at play to multiple people who suggested what bad policies are causing the economic issues and then didn’t provide them until called out.
You’re running partisan defense when you act that way
Obsidian is great but it's a productivity trap for me. The last time I got into it I went too far in designing the 'perfect' PKM system, while not actually using it all that much. Turns out I just really like designing systems. =p
