> I can both deny you the means to defend yourself, forcing you to rely on me for protection. That's the definition of a protection racket.

The US didn't deny Europe the means to defend itself. Europe chose not to build those means because it was cheaper to rely on the US. These were domestic political choices made by governments whose voters preferred social programs over defense budgets. A protection racket requires coercion; what the EU received is much closer to a subsidy.

> This is revisionist history at best. The US has done their best to undermine and dismantle European social programs.

Can you cite a specific example? The US has broadly pushed for capitalist markets or free trade via policy, but "done their best to undermine and dismantle European social programs" is a very strong claim without evidence. Norway's sovereign wealth fund being "tolerated" because of strategic positioning is, at best, a conspiracy theory. There has been some tension over Norway divesting in American companies for political reasons, but that's hardly the claim you've made.

> America was protecting Israel's trade routes. Let's be clear. European trade routes largely just rerouted around the Cape of Good Hope.

Rerouting around the Cape added weeks of delay and a high monetary cost to European shipping. Just because European ships could reroute doesn't mean the European economy wasn't significantly impacted. Why did the European trade association publicly beg for more governments to join the operation if the Red Sea shipping was only about Israel?

> You can't both have a protection racket and expect military help

You expect America to adopt a one-way obligation where it provides for the defense of its allies, and receives no help in return? Why wouldn't that deal fall apart?

> Yes, literally nobody wanted the US and Israel to launch an unnecessary, unprovoked and ill-planned war on Iran

You can disagree with the decision to strike Iran. But when Iran retaliates by launching missiles and drones into 12 different countries (11 of which had not participated in the initial strikes against Iran in any way), the question of whether allies will support defensive operations is separate from whether they endorsed the initial strikes.

> Not a single defense analyst would even seriously consider such a prospect

No country would seriously consider it a prospect because the entire might of the US Armed Forces would immediately engage anyone who tried. This despite the fact that Canada has anemic defense spending, a large arctic border with Russia, and strategic assets I'm sure Russia would love to have.

I believe those people are being a bit silly, and their position probably comes from a strong dislike of Trump as a person, and not a genuine belief.

Russia annexed a warm-water port and then shortly after attempted to incorporate Ukraine as part of a plan to remake the USSR. The only thing keeping China from taking Taiwan is the United States.

The US has no desire to annex Canada, and it also has no need to. If Canada proposed statehood or even a territory agreement with the US, I genuinely don't think it would even pass a vote.

The "protection racket", in particular, is very dishonest. The US has spent 3-4% of GDP on defense for decades, outspending the rest of NATO combined, while the majority of NATO members continuously fail to meet their monetary contributions. Most of America's allies would not be able to fund their generous social programs if the majority of their military capabilities weren't directly tied to the implied threat of the US military interceding.

America's allies haven't necessarily been that reliable for us either.

During Operation Prosperity Guardian, Houthis started attacking commercial shipping vessels in the Red Sea, directly threatening European trade routes, and the US could barely get token naval contributions from allies. The US deployed an entire carrier strike group while Norway sent ten staff officers, the Netherlands sent two, and Finland sent two soldiers. France, Italy, and Spain refused to participate; Denmark contributed a single staff officer while being one of the primary beneficiaries of the US naval protection.

With Operation Epic Fury, the US asked to use jointly operated bases for staging, and Spain banned the US and then demanded that the American tanker aircraft leave. The UK refused to provide any support until drones hit a UK base in Cyprus, and even then, their mobilization was extremely slow. They weren't even able to deploy their carrier, the HMS Prince of Wales, without getting an escort from France. Canada praised the removal of Iran's nuclear capabilities, while providing no support and heavily criticizing the operation itself.

Can we actually be clear on "reliability"? There is not a single defense analyst in the world who seriously believes the US wouldn't IMMEDIATELY defend Canada if Russia launched an offense against them. The unreliability comes from trade policy (which I think is mostly dumb, but is also very much not a one-way action), hesitancy to fund Ukraine at levels that aren't being matched by NATO allies, and Trump's blustering about "adding a 51st state" (no one seriously believes the US is going to annex Canada).

America will continue to act as a deterrent against military action for her allies, and said allies will still not have to commit to the spending that would be required to field a military that is actually a near-peer to China or Russia.

Having said all of that, I 100% support America's allies building out their own cloud infrastructure and bringing defense R&D and manufacturing back locally. Israel has been moving to cut direct dependency on the US and instead acts as a partner in new joint defense capabilities. I think a similar strategy for Canada and Europe would be best for all.

I'm honestly not sure how practical an EU counterpart to Starshield is, but maybe a partnership with SpaceX would allow them to more realistically diversify while the EU builds up its space capabilities.

I wonder how well a similar approach would work with elixir + python. Elixir obviously has very easy process isolation, but I think you'd be stuck using a NIF approach for Elixir, which probably removes any reason to try capn'n proto over just protobufs?

My original thought was to spin up SQLite databases as needed because they are super lightweight, well-tested, and supported by almost every programming language. If you want to set up an agent in another programming language via MCP, but you still want to be able to access the agent memory directly, you can use the same schema in a SQLite database.

I may end up using mnesia for more metadata or system-oriented data storage though. It's very well designed imo.

But one of the biggest reasons has just been the really nice integration with DuckDB. I can query all of the SQLite databases persisted in a directory and aggregate some metadata really easily.

You can safely swap out agents without redeploying the application, the concurrency is way below the scale BEAM was built for, and creating stateful or ephemeral agents is incredibly easy.

My plan is to set up a base agent in Python, Typescript, and Rust using MCP servers to allow users to write more complex agents in their preferred programming language too.

I didn't want to lose the work I had done, and I knew it would be a pain to do it manually with git. The model did a fantastic job of iterating through the git commits and deciding what to put into each branch. It got everything right except for a single test that I was able to easily move to the correct branch myself.

Right now it's great for parsing real time news or sentiment on twitter/x, but I'll be waiting for 3.5 before I setup the api.