beisner

They’re unwilling to pay for fast mode because of the current step function price increase once you hit your quota. It’s a psychological effect. Because most shops I know in the US currently paying $125/mo per seat for Claude would happily - HAPPILY - pay 2x, and begrudgingly pay 10x that amount for the same service. If fast mode was priced 25% or 50% more they’d happily pay for that too. But it’s just not priced that way currently with weird growth subsidization & psychology.

Seems like a pretty obvious effect no? There’s a major power imbalance between professor and student. Not sure how much it would extend outside of the classroom.

Closing certainty is a major reason people will take all-cash offers over financed ones. Even if you're pre-approved for a loan.

Unfortunately the problem is literally the way the government is structured from an electoral + mathematical perspective. Particularly heinous failure mode is polarization, which has been the norm for 50+ years (really started after Vietnam). Biased towards inaction and status quo structurally. The last sustainably unifying event was WWII, which doesn’t bode well.

The randomness is whether the committee reading your essays read them before or after lunch, or if something you wrote reminded them of their first romance, etc. etc. etc. The scores may not be "random" in the truest sense of the word, but the latent state that determine them is unknowable a priori and therefore the scoring ends up being highly stochastic.

I think OP is speaking about people who have 5x or more than their current salary in equities, where even a 50% temporary drop in assets will have no meaningful effect in lifestyle (and a long-term 50% drop in S&P would be apocalyptic, bigger fish to fry).

Intellectual property is theft

On Linux I’ve done this by pinning processes to a certain range of CPU cores, and the scheduler will just keep one core free or something. Which allows whatever I need in terms of management to execute on that one core, including SSH orUI.

For events specifically, my cohort (somewhere between Gen-Z and Millenial) have moved event organizing entirely to Partiful, which I've found to be far superior to Facebook Events. Doesn't help with group posts though.

Nothing in this article (nor its citations) claim that the product is being discontinued. Ramping down production of this version is not the same as discontinuing - and even that probably doesn't have a huge impact on whatever long-term vision Apple has for this product line.

Wonks and experienced New Yorkers aren’t dumb. I too refused any apartment that wasn’t “No Fee”. But only a tiny fraction of the housing stock - especially very few “cheaper” places - are No Fee. And it is genuinely hard to not price anchor on the listed price - especially if you don’t know how long you’re going to live there (aka how much to amortize vs a No Fee)

This is a win for price competition - the "broker fee" paid by the renter is a classic example of principal agent problems / information asymmetry.

First, the service is being provided to the landlord (listing, tours, etc.), not the client, for all listings these days (I don't know any young person who has ever used a renter's agent, except maybe if it's provided in a relocation package). The renter has no choice in which broker to use to find/transact w/ the property, so there's very little price pressure for these broker fees.

Second the information asymmetry - the terms of the fee are completely opaque in the listings, and are not disclosed basically until signing unless you press brokers earlier. So there's basically no competitive pressure pushing these fees down, since it's basically a "junk fee" from a user experience perspective tacked on at the very end (and not listed on listings), and the landlord - who IS in a position to negotiate on price - doesn't care.

I don't buy the argument that there will be some long-term price hike in rents as a result of this decision - people who rent for 1-5 years already are paying a MASSIVE "net effective" premium for having an additional month's rent tacked on up front - but also it strongly incentivizes tenant retention (e.g. by being more responsive, keeping prices lower, etc.), because the landlord does not want to have to eat a broker's fee next listing.

If nothing else, the uncut video of the robot autonomously folding laundry for 10 minutes is incredibly impressive. Regardless of current robustness, or how long it took to get that specific clip, or how engineered the prompts are, this is no small feat.

If you’re only doing vacation travel planning, sure. But there’s a long tail of advanced functionality used across all kinds of industries (with plugins upon plugins) that are most certainly not even close to being supported by any of the options proposed.

Yeah when you have way more applicants than you can possibly look at, the usual strategy is to just randomly sample 10% to review, and throw away the remaining 90% of resumes without even looking at them. This makes intuitive sense, as you do not want to hire unlucky people.

This sounds like a joke, and it is. But this is actually what happens in practice (with less rigorous random disposal).

Turboprop is different from jet

Things actually changed in 2018 with some implementation details of https://en.wikipedia.org/wiki/First_Step_Act. Parole is not given, true; but you can earn a bunch of credit days for participating in non-recidivism activities, which - when combined with good behavior - can amount to leaving prison 30-40% early. Not sure about the details after (house arrest or halfway-house may be somewhere in that mix), but not as grim at that point.

I hate that this is such a common thought process.

In most teams, not everyone is on the same page:

- There are people with varying levels of experience, some may not currently have enough background to fully understand all the aspects that a plan addresses.

- Many/most people have the tendency to let their various biases creep into team-level decision-making. For instance, if someone had a bad experience on a previous team using technology X, they are less likely to support using X today independent of X's suitability for the current situation. And people tend to have a massive bias towards their own ideas rather than others' - so when two competing ideas are presented, one's own ideas are often preferred to others' ideas.

- Especially in larger organizations, there may be political motivations for pursuing one plan over another plan - both for management and for workers. "Nobody got fired for buying IBM" and "resume-driven development" and "launches get promo more than landings" are a few examples that come to mind that I've seen.

Unless you're on a golden A-team of contributors who are all incredible and on the same wavelength (this does happen sometimes! if so, cherish it!), you're going to have a mix of people with a mix of motivations+biases+experiences, incapable of reaching consensus around the "true" best plan in a reasonable amount of time.

Making decisions in a reasonable amount of time is CRUCIAL to team performance. Obviously, sometimes you make a bad decision quickly when you might have arrived at a good decision with more deliberation+debate+waiting for consensus. But in my experience, the benefit of avoiding these "bad" decisions from time to time is dramatically outweighed by a team's ability to decide very on things very quickly and commit to a plan (and therefore execute).

Famously, section 5.11 of the CIA's Simple Sabotage Field Manual [1] describes some pretty timeless ways to ruin an organization's productivity - all kinds of behavior that are kind of indistinguishable from the behavior I've seen arise from "disagreeing and not committing" in organizations.

[1] https://www.cia.gov/static/5c875f3ec660e092cf893f60b4a288df/...

This doesn’t necessarily follow - it may be that the average potential buyer wants the same amount of house as before, but due to supply issues the lower end of potential buyers can’t afford to buy anything at all. And size doesn’t scale linearly with price - it’s usually sublinear. So it could be that the only homes that get built/sold are larger because nobody is building homes for middle-lower income price ranges.

Doesn't work.

if you have a reproducible build, then the notion of "software was built on date by user" is kind of useless information, no? Because it does not matter - if you can verify that a specific git hash of a codebase results in a particular binary through reproducible builds, a malicious adversary could have built it yesterday and given it to me and i can be almost surely confident (barring hash-collisions...) it's identical to a known trusted team member building it.

Having information about which git has was used, as well as the time it was published, is part of the source distribution so an output can contain references to these inputs and still be deterministic w.r.t. those inputs.

If you REALLY want to know when/who built something, you could add in an auxiliary source file which contains that information, which is required to build. Which is essentially what compilers which leverage current time do anyway, it's just implicit.

I think it's probably because it's extremely hard to understand what *content* keeps users subscribing, and "hours spent watching this content" is maybe the lowest-hanging (or only?) "clean" metric you can use to determine what content to keep, how much to spend on content, etc. I watch a smattering of shows on netflix and so keep a subscription, but I don't watch very much. How the heck are they supposed to know whether a piece of content will keep me subscribing? And how much should they spend to produce/acquire that content? It's a very difficult supply-side question.

Annoying that it's not Thunderbolt 4...

One dimension of improvement could be “did the piece I finished/delivered realize my internal vision of this piece?”. In the sense that when you are executing a work you might not feel it captures what you set out to achieve, and may know/not know what was needed to rectify that. Kind of like the old Ira Glass quote on taste vs skill.

I would but https://twitter.com/nvidia doesn't have an well-supported account on Wayland yet :(

OP means something more like "a security which appreciates on average faster than inflation", AKA an asset I can purchase and hold and sell later for more than I paid for it+improvements in real terms.

Is it possible that GPT-4 already had this new optimization somewhere in its training set? The optimizing patch DeepMind published on has been floating around for several months at this point…

Edit: I found the merges for sorting (Jan 2022) [1] and hashing (Jan 2023) [2]. Both of these are very plausibly in the training set for GPT-4, which was frozen sometime in March 2023.

[1] https://reviews.llvm.org/D118029

[2] https://github.com/abseil/abseil-cpp/commit/74eee2aff683cc7d...

That was Uber.

Yeah the USDS is one of the only romantic/bright spots for good application of tech in the US government.

OP is making a dumb libertarian argument basically saying "there's nothing stopping people from doing what they want except the threat of force/violence" (which is trivially true, and the mechanism by which any group of people - even democratic societies - imposes their will on insiders or outsiders), and "do you really want to use THE THREAT OF VIOLENCE to prevent me from STARTING A BANK???" as if it's different from the enforcement of any law. OP questions the use of "should" uncharitably, because the original comment's unstated assumption is clear - society "should" be set up to tradeoff freedom and minimizing collective harm. Violence is not the only form of harm; unregulated banking can obviously cause harm in the society (starvation was pretty harmful in the US during the Great Depression), and so it's valid to question whether or not society should consider how the benefits of private banking (with any amount of regulation) trade off with a public alternative.