jessriedel

> My understanding is that smaller rouge plants do not form in isolation and are always the product of ejection.

Wikipedia says "Rogue planets may originate from planetary systems in which they are formed and later ejected, or they can also form on their own, outside a planetary system."

> The definition is that a brown dwarf forms like a star through gravitational collapse within a protostellar nebula

Do you have a cite on that? (It conflicts with Wikipedia, for example.) In particular, I suspect that many people read something that describes how a certain type of star is thought to typically form and then interpret that as a definition.

I am questioning whether the article is accurately describing the scientific nomenclature. The author of course wants to say things are being “upended”, but I suspect the author just didn’t understand, or is inaccurately describing, the previous state of affairs.

> But distinguishing between smaller brown dwarfs and planets requires looking at how they form. Even though they can end up about as massive as 10 Jupiters, planets always arise around a star, from its surrounding disk of gas and dust.

> In contrast, stars, including brown dwarfs, form on their own within giant collapsing clouds of gas.

Is this really the standard terminology? It’s not how I remember it, and it doesn’t make much sense. There are tons of binary star systems, and while a some are three-body capture events, aren’t most formed from the same gas cloud? (I.e., not “on their own”.) Likewise, rogue planets (i.e., not bound to a star) can be formed in a stellar system and be ejected, but can’t they also form on their own, e.g., a dust cloud with less than enough total mass to form a star? Surely you wouldn’t call a sub-Jupiter-mass body a “brown dwarf” just because it formed in isolation?

My point is that it’s largely a semantics dispute, not a scientific one.

My point isn’t that these can’t be interpreted correctly, it’s that these framings teach us nothing new because they have been repeated a thousand million times and this article does not attempt to go beyond the superficial cliche.

This article managed to hit two classic science journalism cliches in just the first few paragraphs.

(1) “Someone hypothesizing a very dramatic theory with weak evidence was considered wrong by most colleagues but later vindicated when strong evidence emerged”. (No mention of thousands of other dramatic hypotheses that turned out wrong.)

(2) “You may have heard in unsophisticated popularization that [philosophical claim ultimately hinging on semantic distinction] was false, but really it’s true [assuming my preferred semantics]”.

Aren't we all tired of this yet? Aren’t science journalists embarrassed by this stuff?

Ok. Personally, I dispute that the designers don’t care about efficiency. I think it’s just expensive.

The hypothetical is only interesting if we draw a distinction between easy and hard tasks. (“AI can do anything a human can do behind a keyboard” is AGI complete.) For this hypothetical I asserted that performance optimization and porting were automatable, but that doesn’t include any feature.

If you think there’s a more interesting/realistic/well-defined assumption, please suggest it!

The assumption of the hypothetical is not that they can deliver any feature, which I consider (maybe you disagree) more open-ended and less automatable than porting and optimization.

If you think that’s an unrealistic assumption, please feel free to propose another!

Because features require more human overhead to understand and manage than porting/optimizing does, by assumption of this hypothetical. Product managers and high-level architects aren’t free.

Microsoft would hire a ton more engineers and look very different if engineer price went to ~zero. So what does that look like? If you can make an app snappy with $100 in GPU time, why wouldn’t they?

This is often explained as “cycles are cheap, programmers are expensive”. For the sake of argument, suppose LLMs are able to replace programmers for “laborious but straightforward work” like writing native versions of your app for every platform, and optimizing the hell out of it.

Do we then expect all the apps we work with to become snappy again? Or are there red-queen-like dynamics that will prevent it?

There are tons of effective QFTs that can be used very accurately to predict experiments below a certain (high) energy threshold, but which are not renormalizable. This is well understood. It means that the theory needs to be augmented to understand the higher energies, but does not at all mean the more fundamental (augmented) theory has to be radically different, such as classical. Everyone already agrees that the non-renormalizability of naive low-energy quantum gravity means something must be added.

That's not an explanation. You could say that about a thousand things in McDonald's business. They are all a relatively small part, but in fact McDonald's has an army of people optimizing them all, and they are overall very successful.

Why do you think they would be able to abuse any of the operators? There are dozens and dozens of fast food franchises in the US, all in cut-throat competition. You may have heard of McDonald’s, Burger King, Wendy’s, Subway, KFC, Chick-fil-A, Domino’s, Pizza Hut, Dunkin’, Sonic Drive-In, Popeyes, Little Caesars, Arby’s, Chipotle, Dairy Queen, Jack in the Box, Hardee’s, Panda Express, Jimmy John’s, Five Guys, Wingstop, Culver’s, Zaxby’s, Raising Cane’s, Whataburger, Bojangles, Shake Shack, El Pollo Loco, Firehouse Subs, Jersey Mike’s, Del Taco, Checkers, Rally’s, Church’s Chicken, Moe’s Southwest Grill, Qdoba, Captain D’s, Freddy’s, Tim Hortons, Smoothie King, Blaze Pizza, Baskin-Robbins, White Castle, Portillo’s, Noodles & Company, Schlotzsky’s, Tropical Smoothie Cafe, Auntie Anne’s, Marco’s Pizza, Boston Market, Smashburger, Fuddruckers, WingStreet, Krystal, Papa Murphy’s, Hungry Howie’s, The Habit Burger Grill, Jamba, Nathan’s Famous, Steak ’n Shake, Waffle House, Big Boy, Pizza Ranch, Cook Out.

I don't understand the last sentence. If the machines are frequently broken, that damages the Macdonald's brand in the consumer's eyes. And if the franchisee's are paying unnecessary costs, making a Macdonald's franchise less lucrative for the owner-operator, that will lead to fewer franchises renewals and new franchises in the future.

Shouldn’t we have a non-negligible prior on possible cancer risks at all timescales? That doesn’t mean we don’t approve drugs without 80-year clinical trials, but I think it’s a non-negligible risk that has to be accepted.

For instance, my impression is that lung cancer rates only really become strongly pronounced — in the sense of being statistically detectable in the small population sizes you find in clinical studies, rather than the entire US population — after like 15 years of smoking. (Correct me if I’m wrong.)

Rules degenerating into infinite whack-a-mole is a strong (though inconclusive) signal a mistake is being made. "Let's ban rent increases". "Whoops, now all the landlords are slacking on property maintenance; let's mandate maintenance." "Whoops, now all the landlords have stopped making improvements; let's let them increase rents X% when they spend at least $Y on improvements." "Whoops,..."

So you end up in some new equilibrium. Maybe that equilibrium is better, maybe it's worse, but it's simply not true that it's always better to do something rather than nothing, and pointing out the loopholes in the rules is valid criticism.

I think it’s just a personal quirk

Also, ~every physics paper since 1993 is on the arXiv. The same is true for math and CS with later cutoffs.

I think studying this stuff is always going to seem mysterious unless you account for the concept of fashion in science. Specifically what I mean is that two papers (or ideas or approaches) X and Y can have equal “objective scientific merit” but X is more popular than Y because of random initial conditions (e.g., a famous researcher happened upon X first and started mentioning it in their talks) that are self-reinforcing. The root cause of this phenomenon is that most/all researchers can’t justify what they work on from first principles; for both good and bad reasons, they ultimately rely on the wisdom of the crowd to make choices about what to study and cite. This naturally leads to big “flips” when a critical mass of people realize that Y is better than X, and then suddenly everyone switches en mass.

I’m responding to the suggestion that being on the non-engineering side explains his opinion.

>> Almost everybody I know uses ChatGPT at least semi-regularly for all kinds of things.

> Ed's perspective is from a nonengineering side, not the typical Hacker News user or LLM API consumer.

ChatGPT has 200M weekly active users.

The booster does not travel more than ~50 miles.

A major difference is that F9 (landing on a wide flat pad) had a quite wide acceptable horizontal error, 10 meters or more, whereas I think this (landing between two chopsticks) needs like ~1 meter accuracy in the radial direction.

The booster is always returning to the general vicinity of the launch tower (either the tower itself or a barge). It isn’t used anywhere unimproved, and in particular is not used on Mars. So what scenario would it be helpful to be able to land the booster on a flat concrete pad?

Thx!

What's the highest resolution image of the moon taken from Earth? Presumably some stitched together image from one of the large telescopes?

> Some more education is still necessary, though. Certain foods can go months or years past their "best by" date and still be perfectly safe - even if they might taste a little funky.

Aren’t safety restrictions the “use by” date?

> Switching to language that either says “Best if Used By” or “Use By” will also help minimize ambiguity. The former will establish consistent wording for advising when an item is less fresh (but still okay to eat), while the latter designates food that should no longer be consumed due to safety concerns.

I’m specifically talking about the situation where you lose almost all the cases. When that happens, it’s not called “clearing up the law”, it’s called frivolous lawsuits.