I prefer “rare” to “well-done” — in steak, and in life.
Algorithms tend to optimize us toward well-being as “well-done”: predictable, consistent, uniformly cooked.
Safe, measurable, repeatable.
But human experience is closer to “rare”:
uneven, risky, asymmetric, and still alive.
The parts that matter most are often the ones that don’t fit cleanly into metrics.
If everything becomes optimized, nothing remains interesting.
And more importantly, we risk replacing well-being with the monitoring of well-being.
When a life is constantly optimized, scored, nudged, and corrected,
it gradually stops being a life that is actually experienced.
Thank you for this deeply revealing take. I think this is the dynamic at the core of what matters here. Reminds me of Dostoevsky's take on what people really want - here's an interesting short piece that direction.
If a model eventually scores perfectly on every benchmark yet ends up practically useless, what’s the next step?
Benchmarks measure competence inside a predefined problem space,
but real scientific and engineering work isn’t bounded — it keeps changing underneath you.
At some point we don’t just need a system that knows how to solve problems in theory;
we need one that can actually do something with that ability.
The equivalent of making the coffee when we want coffee,
not just getting a perfect score on a coffee-theory exam.
multiple satellite operators are coming on line. what are the odds all of them coordinate to shut down in one region invalidating using the other providers as fail over?
I might be mistaken, but as far as I know there is currently no other LEO broadband provider that is meaningfully comparable at a global scale.
Starlink is often treated as the reference point not because it is perfect or fully resilient,
but because there is no second network at a similar scale that could realistically serve as a failover today.
If we imagine a hypothetical future where three mature operators exist, then yes — absent coordinated political or geopolitical action, at least one network might remain online.
However, even that surviving operator would not necessarily provide full coverage of the affected region.
Global redundancy is extremely hard in practice, because maintaining continuous, worldwide LEO coverage is not free — it requires massive capex and opex, ground stations, regulatory permissions, and local political approval.
True worldwide failover remains more of a theoretical construct than an operational reality.
That particular section I have to Wikipedia article seems to have gone through a bunch of anonymous edits back and forth around the content of this citation
Probably because it's not actually a truthful characterization of what happened! I know it's popular to find every possible reason to bag on Musk, but you don't need to resort to disinformation to do it.
They didn't, and you're again repeating misinformation as fact.
What happened was that they refused to turn access on for the Crimean region, which is not the same as "cutting off Ukrainian access".
I understand nuance is hard to grasp sometimes, but if you're going to continue to conflate the two things, I can only chalk it up to a desire to deceive.
Can't talk for the USA, but it's widely acknowledged that the spread of broadband in Europe was driven by P2P and tools like Emule/eDonkey or BitTorrent.
We need some similar killer application for satellite connectivity and mesh networking. Something that makes the technology so requested and so ubiquitous in such a short time that it couldn't be banned even if they tried.
In Tanzania they went around to hotels during the ban to make sure they didn't have starlink. It's illegal here but many have it. During that time some enterprising individuals charged tourists to access theirs.
The beam forming used by Starlink (and Starshield) is highly resistant to jamming. But Starlink doesn't offer service in some countries. And the ground terminals can be detected.
It might function as a kind of cogeneration-style buffer, but CO₂ still gets emitted in manufacturing and maintenance — and I’m not sure the volumetric efficiency is all that compelling.
Still, if we ever end up with rows of these giant “balloons,” the landscape might look unexpectedly futuristic.
Algorithms tend to optimize us toward well-being as “well-done”: predictable, consistent, uniformly cooked. Safe, measurable, repeatable.
But human experience is closer to “rare”: uneven, risky, asymmetric, and still alive. The parts that matter most are often the ones that don’t fit cleanly into metrics.
If everything becomes optimized, nothing remains interesting. And more importantly, we risk replacing well-being with the monitoring of well-being.
When a life is constantly optimized, scored, nudged, and corrected, it gradually stops being a life that is actually experienced.