Trump already tried to overturn one election he lost. On November 3, he’s likely to lose another one, badly. At this point it is obvious to almost everybody that the Republican Party, and thus Trump, is going to get smashed in the midterms. Chances are that both the House and the Senate will be flipped. A year ago, it was mostly unthinkable that the Democrats could flip the Senate. Although it would be a massive loss, even for a midterm, in this post I want to make the point that this is not just another election.

Trump doesn’t like to lose

I don’t think Trump would take this loss lightly, especially after a potential landslide. I think Trump will challenge, litigate and threaten to get the result of the election overturned, at least in key states. Some of the litigation has already started.

When I mention this to folks, so far everyone replies with skepticism: they don’t think that could happen, or that it’s even possible. To me this is a clear example of normalcy bias, and it’s quite easy to show why: Trump has already tried to overturn an election that he lost, in 2020. To this day, he doesn’t admit that he lost that election, and neither do key members of his cabinet. That alone should have been a massive red flag against voting for him in 2024, but here we are anyway and it’s about to happen again.

There is speculation that the administration will try to influence the elections by using force, but as this excellent piece from Nate Silver’s newsletter describes, it may not even be possible to do at scale, although they might pull it off in a specific key race. Even if it doesn’t happen, the fact that we’re talking about it at all says a lot about how unstable things could get.

Some weeks ago my post “How the AI bubble bursts” reached #1 in Hacker News and stayed there for about 3 hours, plus 1 hour on the front page until it ultimately dropped. This came as a total surprise, I expected to get 4 points, like my previous post.

It was certainly an emotional rollercoaster to get that validation from a community I’ve been following since I first got to the internet. It was also hard to not make this my whole personality. It’s crazy to think that, for 4 hours, my writing was the most-read thing in Silicon Valley, that people I admire, from companies I admire, even maybe a billionaire, read my article and got an idea out of my post. It’s almost certain that at least someone from a major AI lab read it.

The aftermath

The post ended up with 372 upvotes and 532 comments, was opened by 16K people the day it was released, and very interestingly, by 4K more the rest of the week. My Cloudflare Pages setup for this blog was able to handle this without blinking, which was a huge relief. As soon as it got to the top, Google Analytics showed me 6K people online. Surreal.

Something I didn’t know is that after a post makes it to the homepage, Hacker News drops the nofollow tag, so Google is able to recognize it as a full link. The article also gets mentioned in the front page archive for that day. It was picked up by the Financial Times as a reading list for the day (also with a full link), and also linked in a dozen Hacker News aggregators. It got backlinks from other posts where they mentioned it, a bunch of shares on subreddits. It got shared about a dozen times on Twitter and it was featured in an AI-generated podcast.

The backlinks gave a boost to the Google rankings, this blog was started just 3 months ago after all, but search traffic dropped back to baseline quite quickly. I thought it was going to push Google to index parts of my site that it was ignoring, but so far that didn’t happen.

Given the numbers, it yielded a remarkably low long-term engagement: only 12 people left their email to subscribe to the blog and I got a grand total of 2 new Twitter/X followers.

But some people reached out on LinkedIn, and even people I knew congratulated me on making it to the top without me telling them I was there, I didn’t really think that many people read HN daily. Some people told me they saw the post shared in group chats.

Most of the traffic came from the US, particularly San Francisco as you’d expect. Second came the UK, Germany and Canada. I think the Europe traffic is significant given that it was posted in a more East-Coast/Europe time-friendly (1pm CET). Even then, San Francisco did 3x more views than New York.

It took me 9 years to come back to blogging, and I used maintaining the old WordPress setup as an excuse the whole time. Turns out the fix was free. The previous version was a WordPress running on a VPS on Digital Ocean. I could have, in theory, just start writing there and hitting post. But the WordPress version was getting more outdated every day and the plugins I was using were also getting deprecated, so I’d have to migrate. The blog was down every couple of years because of issues in the VPS.

Beyond the operational headaches, there were deeper problems too. One of the key features I wanted to support was multi-language, and getting that out of the box is not supported by WordPress by default. Switching the plugin would mean having to make sure the URLs are kept the same and migrate all the data to the new one. Plus, I needed to make sure it ran, and that I didn’t get hacked, as WordPress has gotten many times. The setup with the VPS would have to be maintained, and I set that up when I was just getting started with software, so I’d probably have to rebuild it from scratch. In case one of my posts was a hit, I’d probably go down or with heavily degraded performance. There are also serious concerns everyone should have about WordPress (the Matt meltdown).

Alternatives to self-hosting

Of course, one-click WordPress-hostings exist, but that’d mean I need to pay money monthly and also risk it becoming expensive overnight if one post gets significant traffic. Migrating to them would also not be trivial if I wanted my old blog to have the same setup, and more importantly, the URLs. I could also host in something like Substack or similar platforms, but that’d mean I don’t really have full control of my website. I don’t have anything against Substack, I even own some shares of it, but I would rather have full control of my little internet corner, where I can write anything I want and not get banned or demonetized. I don’t want to have any external incentive influence what I write.

I also thought about using AI to create a custom blog code for me and host it in something like Vercel, but that felt like more of a passion project rather than something that’d make my life easier.

We are sleepwalking into the worst energy crisis since the 1970s. The Strait of Hormuz has been closed for 34 days now, and with it the world’s supply of 20% of oil and gas, 30% of fertilizer, 33% of helium and 10% of aluminium, among other critical materials.

Brent Oil is already at all-time high, but we haven’t priced in the effects nor have governments reacted. Shortages and high prices around the world are likely to come next. Restrictions have already started in some countries. To give perspective of the depth of the situation, the US had to rush to lift sanctions on Russian, and even more shockingly, Iranian oil.

Our Covid moment

This is just like covid in January 2020. But this time it’s not a novel virus that we don’t know if it’ll kill us all and spread like a plague, or just a flu that’ll go away. This time we have the rare benefit that we know what’s coming: the shortage of the fuel we’ve relied on for 100 years, and many other key raw materials. I think people will start to react when leaders around the world start to issue warnings and limitations to lifestyle, as we’re already seeing in New Zealand and Australia. It is our own responsibility to be prepared.

But it also feels opposite in a way. During the pandemic oil traded negative because there was such demand destruction, now high prices are the ones pushing the destruction.

The catalysts for a crash are already laid out, and it can happen sooner than most expect. AI is here to stay. If used right, chances are it will make us all more productive. That, on the other hand, does not mean it will be a good investment.

Big tech doesn’t need to win, just outspend

Magnificent 7 companies are increasing capex to their biggest ever to differentiate their tech from each other and the big AI labs, but the key realization is that they don’t have to spend it to win. It’s a defensive move for them, if they commit $50B, OpenAI and Anthropic need to go raise $100B each to stay competitive, which makes them reliant on investors’ money. As the numbers get bigger, the amount of funds that can write checks of the size required to fill such amounts gets smaller. And many of them are now getting bombed in the Gulf.

This is the reason there’s a push for IPOs, it’s because it’s the only option left to keep the funding coming.

Taking this into account, Google is extremely well positioned to weather the storm. When they announce capex expenditure, they don’t spend it overnight. They can simply deploy month by month until their competitors struggle to raise and get forced to capitulate. At that point they can just ramp down the spending and declare victory in a cornered market. They don’t need capex, they just need to make it very clear for everyone that nobody can outspend them. It is hard to picture as numbers get so big, but Alphabet (Google’s parent) is ten times more valuable than the biggest military company .

This also has a great implication for the Mag 7, especially Google: their capex will be a lot smaller in practice than projected, and as investors hate to see high capex in tech, the market will probably reward that if it materializes.