LEO is not the place to worry about Kessler syndrome.
Mostly, Kessler syndrome isn’t something to worry about at all; there are just a lot of orbital planes available. But in LEO, the mechanics don’t even apply.
For agentic work, you just cache the prefill kv cache of the relevant system prompts. TTFT is a little slow (10s of seconds, oh no!) the first time you boot up a new harness.
I keep hearing people make this claim that TTFT is a problem, and… it just isn’t, if you’re running oMLX.
Folks in my camp keep saying this, and folks in your camp keep beating a drum we tell you isn’t resonating. Not sure why I keep bothering to argue; you can’t buy a high RAM Mac Studio like mine anymore.
Kids come out as a person, with strong opinions and desires. You can shave off some rough edges, and maybe bend a few branches of their experience.
But if you present a kid with the opportunity to read, and they read, you can’t take much credit. That’s just who they are. Others are given the opportunity and don’t.
You can fail to provide the opportunity, but after that, it’s pretty much up to the kid.
As a life long reader, on my own, and to my kid, including many a night time baby -> toddler -> easy chapter -> harder chapter read, my kid doesn’t read books. Certainly competent to do so, but just doesn’t. Possibly we could have continued to deny access to Netflix until later (it was 1 hour a week until about 10). No YouTube allowed. Still, didn’t read. Other kids do, and I’m jealous.
Before I drove an EV, I drove a 50 mpg Prius. At California prices of ~$4/gallon, that’s $.08/mile.
My post wildfires NEM2 off peak rate for electricity is $.40/kWh. My Bolt gets 4.5 miles/kWh. That’s $.1125/mile.
If I were driving a Tesla it would be worse (my wife’s Tesla lies mercilessly about its range when full; it’s like Elon Musk recapitulates himself; real world it gets about 3.5 miles/kWh), and if I drove a Rivian it would be MUCH worse.
So, in California, it isn’t true at all (mostly because rate payers are funding PG&E’s liability) that the most efficient EVs are cheaper than a good mileage gas car. No where near a 2x advantage (it was better, but not nearly 50%, when I bought it, more like 90% of the gas cost). At no point has it ever been close to 50% cheaper for fuel in California (which, as it happens, sells by far the most EVs).
Generally speaking, I think EV proponents (like me!) should spend a lot less time promoting “it’s cheaper”. It is, in practice, cheaper, because maintenance is cheaper. But Americans don’t care about levelized costs, they care about the highest salience variable expenses, and trying to convince them to do otherwise is a losing argument.
As a committed market-socialist, the extent that the HN commentariat has turned into a reflexively nihilistic anti-free-enterprise is a fairly astonishing turn. It’s like my belief system turned into the maximally idiotic version of itself.
“Externality” is thrown about as a term almost completely disconnected from any economic grounding of the term. If you make externality mean “anything I find aesthetically displeasing”, then yeah, sure, billionaires create and benefit from externalities, if your aesthetic is egalitarian comity.
But if you mean “legitimate societal goals, legislated and agreed on by a representative body” are being violated left and right by billionaires, gimme a break.
Go ahead and tax capital gains way more. Ending the estate step up in basis sounds great. Break up the “borrow” part of buy-borrow-die, while you’re at it, and treat encumbrance on capital as a taxable event, we could probably make that work, too, although the middle class might foam at the mouth if that was applied broadly.
But, man. The cynicism, confiscatory and controlling instincts on display are enough to make me upgrade Ayn Rand from “hypocritical nut” to “maybe she was on to something when the general population gets tall poppy syndrome.”
Markets work. There are externalities, but we can, and should, legislate fixes for social goals that we actually agree on. But stiflingly heavy regulation is really bad for incentivizing creation of new knowledge and wealth. You can still believe in caring about people, and building (incentive aligned) social safety nets without destroying people’s incentive, and thus, because intellectual capital formation depends heavily on network effects, people’s ability, to create many kinds of value in the world.
Actual socialists recognize that capital is incredibly useful, and incredibly valuable. Leveraging capital is incredibly beneficial to the world. Pretending that the people leveraging that capital are somehow guilty of an original sin just by leveraging capital markets, which is really what these screeds against anyone holding controlling interests in companies they were instrumental in creating, seem to be about, leads down a terrible path.
Demonizing people creating things is petty and unbecoming for a political movement.
The idea that 1:1s with devs adding very little value to the team is… pretty wild.
If you think 1:1s don’t add value, your slice of the reality of what even modestly sized teams need to operate smoothly is so far from my experience I don’t think we’re likely to bridge the divide.
But to make a good faith effort: what is the job you think line managers are supposed to be doing, if not listening to devs, going to meetings you would prefer not to sit through, and writing up carefully documented feedback for the under-performers you seem convinced surround you at every turn?
There are many projects where one shot is the right answer!
But surely you aren’t suggesting literally every software project is composed of one-shot-able building blocks, or that the building blocks never require modifications to previous one-shots?
This is orthogonal to the Interchange Fee Settlement.
The settlement allows stores to decline different *classes* of Visa cards. It was always possible to accept Visa but not MasterCard, etc. What was not previously possible was to query, before the transaction “what will the fees be” and reject cardholders presenting high-fee cards from a network you have a relationship with.
That is now allowed, by consent decree. But so far no one is doing it.
Incidence is very much on customers, but (high interchange fee) credit card users are getting a rebate of most, if not all, of that. It’s the cash users who AREN’T getting a rebate, and thus the incidence is on them (and people using other low-or-zero cashback payment methods).
This used to be true. But the 2025 Interchange Fee Settlement abolished the “Honor All Cards” regime. Perhaps, I know it’s crazy, but… perhaps segmenting the market into extremely high-spending customers, normal pay-every-month-relatively-low-fees, and no-frills, was a smart move by the big issuers? My sense is that alienating big spenders (whose interchange fees tend to be in the 4% range) is just not worth it?
All I can say for sure is no store I’ve ever encountered has operationalized the newfound ability to differentially reject some cards yet. I am starting to see grittier establishments offer 5% cash discounts more frequently than they used to, and I’m always happy to pay cash when they do.
But when there’s no discount, why would I forgo better accounting and 3-5% back in points?
Normies very bad at the concept of statistical calibration. News at 11!
But yes, I agree with you that it's surprising to hear people on Hacker News having the 180 degrees wrong impression that the general population appears to have taken away from the one thing normal people care about polling for: during presidential elections.
Huh? Your conclusion does not follow. A large fraction of the interchange fee is kicked back to customers.
The size of the pie being so much bigger means the issuer’s tolerance for fraud is much larger, but it’s orthogonal to whether there’s actually more fraud. In practice credit cards fraud actually impacting customers is vanishingly rare at this point.
Because adding friction will deter many impulse purchases. Americans use credit cards constantly. The equilibrium would be perturbed in a way very much not advantageous for the credit card issuers if consumers became more cautious about using credit cards.
It’s the same reason credit card issuers are willing to pay Apple a few basis points to participate in Apple Pay: reducing friction has a non-linear impact on propensity to pay.
I skimmed various sections. I found the animations pleasant, the text readable, and the content clearly not slop.
The historical context of the telegraph was interesting, and the treatment of bandwidth vs. latency was thoughtful.
I think it’s too long; I don’t think many people who don’t already know most of this material will read it, but I enjoyed the parts I read. Nice work!