mercutio2

Wow. The reactions here. So negative!

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!

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.

This is one of those questions non-parents ask.

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.

Yup. $1/year is a great price point for avoiding squatters and proving there’s someone paying attention, truly de minimis.

I’m happy with my .xyz domain for my homelab.

Setting up split horizon DNS was more of a pain than I expected, but it works now over headscale.

The pacific northwest’s vast hydro capacity makes it maximally attractive to own an EV. The rest of the country isn’t nearly so lucky.

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.

Yes, but you’re perhaps missing OP’s point?

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.

This is The Way.

I wasn't in the room when it happened, but this is very different than the story told internally about why Apple became allergic to Nvidia.

Arguably more petty. SJ has been dead for almost 15 year now, I imagine the C-suite might get over it at some point.

Cmd-W closes the current *document*. In tabbed apps, the document is the tab.

It is true that Finder is always running, you can’t quit it or kill it.

A 486 in 2003? Pentiums were shipping by the mid-90s, did you just have super old hardware lying around?

I retired my 486 in ‘95 or thereabouts…

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.

oMLX makes prefill effectively instantaneous on a Mac.

Storing an LRU KV Cache of all your conversations both in memory, and on (plenty fast enough) SSD, especially including the fixed agent context every conversation starts with, means we go from "painfully slow" to "faster than using Claude" most of the time. It's kind of shocking this much perf was lying on the ground waiting to be picked up.

Open models are still dumber than leading closed models, especially for editing existing code. But I use it as essentially free "analyze this code, look for problem <x|y|z>" which Claude is happy to do for an enormous amount of consumed tokens.

But speed is no longer a problem. It's pretty awesome over here in unified memory Mac land :)

“Scared” to “take risks”?

This is a bizarre way of saying “if they ship it and it has reliability problems, they know they’re skating on thin ice”.

Apple’s brand has taken a beating (I’m as aghast with the latest macOS as the next nerd), but people love that when Apple ships a product, it generally works and the hardware doesn’t break.

Butterfly keyboards are a terrible stain on the hardware team’s reputation. “Scared” is the wrong word for how these things work.

I was working at Apple and wondering the same thing ;)

Turns out people like them. Not so much the HN crowd, but c’est la vie.

I wonder if the reason that most people don’t agree with me about Antigravity is because you were used to VS Code?

For me, Antigravity is possibly the worst GUI experience I’ve had since Clippy.

It’s completely filled with arcane buttons, prompts that are effectively modal appear in at least 3 different places, it’s constantly… doing stuff, without any reference to where I should focus my attention.

I appreciate Google giving away absurdly generous quantities of tokens for FREE just to get me to use the thing, but I can’t bring myself to, because when I get into the flow of a feature with an LLM, I’m suddenly stuck and can’t figure it out.

It’s like peak Google UI for me.

I used to feel this way, then a week or two ago, after an automatic update, it started hanging. All. The. Time for me. Launching the app now frequently takes 30 seconds before it shows me the “load a git repo” screen.

Speed was its main advantage, before. It has become nearly unusable.

The price of extraordinarily rapid iteration, I suppose.

I'm surprised that there were downvotes. This is an excellent answer, and better interfaces between linguistic definitions of color and physicists' than saying "Rayleigh scattering impacts blue more than red"!

I guess I’ve never tried the pro models, because I’ve used gemini-cli free every day for the last three months or so.

It does eventually finish its quota, but then I just switch to a different Google account (which, amusingly, is what Gemini told me to do).

Happy to consume Google’s free tokens! The free model is a distant third for coding, but it’s fine for leaf node work in a larger project.

It hadn’t occurred to me this was a billing bug.

That would be heartening, if I wasn’t consuming tokens 10x as fast as expected, and they just had attribution bugs.

Do you have references to this being documented as the actual issue, or is this just speculation?

I want to support Anthropic, but with the Codex desktop app *so much better* than Anthropic’s combined with the old “5 back and forths with Opus and your quota is gone”, it’s hard to see going back

Yeah, the generosity of Anthropic is vastly less than OpenAI. Which is, itself, much less than Gemini (I've never paid Google a dime, I get hours of use out of gemini-cli every day). I run out of my weekly quota in 2-3 days, 5-hour quota in ~1 hour. And this is 1-2 tasks at a time, using Sonnet (Opus gets like 3 queries before I've used my quota).

Right now OpenAI is giving away fairly generous free credits to get people to try the macOS Codex client. And... it's quite good! Especially for free.

I've cancelled my Anthropic subscription...

Wealth taxes are very, very different from higher income taxes.

People are mad about buy-borrow-die, so they’re proposing extraordinary new measures.

Personally, I’d just make capital gains taxes apply at the “borrow” stage to actually fix the problem. That would have a host of compliance issues but they’d be localized in the finance industry which already has an army of people figuring out compliance.