sparsely

There's a lot in Solaris (at least in the 1976 movie and the book). In almost all sci-fi which contains aliens, they're easily comprehend-able by humans - they have motivations we can describe in our own language, a physical form that while strange is likely a mashup of earth creature characteristics (after all that's where our vocabulary comes from). Solaris asks what it might be like if this isn't the case, and where the limits of our comprehension are.

> China released its rule for stage 6 light-duty vehicle emissions limits in December 2016, so manufacturers have had 7 years to bring their vehicles into line.

Don't feel too sorry for them if they couldn't read the laws in the country they were selling in.

That would definitely be better, although you're still only capturing part of what's available, or at least a biased sample of what's available. Different form factors come onto the market at different rates - some may never be on the market, or not in an easily accessible manner (sold locally, or via word of mouth, or via private auction etc). I think perhaps that's a distinction that the article fails to make, it's easier than ever to access goods and services from all over the world, but that ease also favours mass market products. If you put as much effort into doing whatever you're trying to do as someone would have pre-internet, you probably have access to at least as much variety as they did.

It's indeed unsurprising that if you look at the designs produced to match a specific context (AirBnB) you'll get a good amount of uniformity, as sellers converge on efficient solutions. If you looked in other contexts (high end apartments for sale in major city, cheap new builds in small towns, mass produced single family homes in another country) you might end up finding more differences.

Yes, that's the one, thanks!

There's a game this made me think of, which I now cannot remember or work out the name of. You had a collection of objects (think coloured plastic triangles, cubes etc.) and one player arranges a few of them into groups, which should all match a secret rule that they have come up with ("more red than blue objects" or "all triangles are stacked on top of cubes"). The other players then have to try and work out the rule, but the rule setter only communicates by giving counter examples if they haven't got the rule right. It was something like that anyway, can anyone remember the name?

The original edition was kind of cheap and plastic feeling, but it's the sort of thing that could be done nowadays with really high quality components.

> I’m a bit concerned about what happens when it freezes. Can the outdoor unit still I operate?

Yes as long as the parts where chosen by someone with a bare minimum of competence. The efficiency will be lower than ideal but still better than electric radiators. Norway has very high heat-pump installation rates.

Indeed. How would you go about scientifically assigning something like "agency" or its lack to any particular entity?

In ordinary language we are perfectly happy to use agency-adjacent words when talking about non-animal entities "I'm hiding from the rain", "That marble on the floor isn't going to get away from me" etc. Obviously in many cases we would also deny them agency, but I think it's hard to be too dogmatic about this unless you're playing some specific language game.

I don't think a logical positivist would have much a problem with the content (although perhaps they wouldn't like the language used). How would you go about verifying that the Mississippi does not have "agency" but a slow worm does? They'd perhaps dismiss it as a meaningless question in the context of scientific enquiry, but fine to discuss in another context (in another "language" as they might put it).

I tried this and whether or not you think it worked probably depends on your level of cynicism

Prompt: In the future please give me two answers. One with a temperature of your choosing, appropriate to the question asked, stating each time, which temperature was used in the response. And the other one with a fixed temperature of 100.

What are some crazy ideas to reduce crime?

~~~~

Answer 1 (Temperature: 30°C): One crazy idea to reduce crime could be to create a virtual reality program that allows individuals to experience the consequences of their criminal actions without actually committing them. This could serve as a deterrent and encourage individuals to make better choices. Another idea could be to implement a reward system for reporting crimes, where individuals are incentivized to report crimes and provide information to law enforcement.

Answer 2 (Temperature: 100°C): Some potential solutions to reduce crime could include increasing funding for education and social programs, providing more job opportunities, improving access to mental health services, and investing in community policing strategies. It's important to address the root causes of crime rather than just trying to address the symptoms.

"Solaris" is an interesting exploration of a situation in which this is maybe not the case.

I believe you are describing something like this: https://mrcool.com/u-shaped-window-a-c/ Never used one myself but they are meant to be much better than the ones with ducts - you don't lose so much efficiency from the ducts or from badly patched up window openings.

I don't have exactly the same problems - our dev stack if pretty old fashioned but works well (dotnet). I don't enjoy it as much either though, most business problems are either easy + tedious to resolve, or involve making a complexity tar pit even worse. Naturally you'll end up spending almost all your time on the latter. Add on top a lot of interpersonal tension from there being a bunch of almost equally good ways of doing any one thing and it's easy to see why programming can stop being fun (while still being a decent profession).

I thought this was well put: > I now understand why there’s a trend of developers who want to go live in a farm or take up woodworking: tough(er) problems, but with less variables & easier to reason about. If the wood breaks, you can see where and can probably guess why. Making a table is a mostly linear set of steps, and the basic tools you use don’t change much throughout the years. There is no invisible ghost that lives in a separate realm (dev environment) that can ruin your work at any time and leave no trace.

There's not many things less starting an implementation and discovering a series of hard to anticipate incompatibilities between your implementation and the rest of the system. I'm sure this happens in other professions, but perhaps it's easier to work around?

Upvote if you identify as T-shaped

> Putting garbage in containers is the kind of thing that sounds like a great idea at first glance

Also at second glance, and a third if you go look at other cities in the USA and around the world. It works great! It solves this problem! NYC should just do it.

I've been to Lunettes pour Tous stores https://lunettespourtous.com/ in the past. It felt revolutionary, because you could walk in, get an eye test, and have your glasses made within 30 minutes at the same store, at a very reasonable price.

It looks like they deliver as well. Hard to describe their price range as they have a million options you can pick as part of the ordering process - easy to end up with something well under 100EUR though without picking all the cheapest options.

I don't really understand how this analogy works in terms of the actual practice of running load tests. You should of course test the full flow, the important part is not implicitly limiting the input volume based on the system performance. I think in your analogy it would be something about people queuing outside without you noticing but am unsure.

Also, to be clear, the bad thing is normally doing closed workload tests, not having a multistep workflow or whatever (https://gatling.io/docs/gatling/reference/current/core/injec...)

This seems to encourage you to run it with a target number of concurrent users, rather than a target RPS from the loaders, which IME can result in difficulty interpreting the results (it also doesn't reflect most use cases, except in some heavily controlled scenarios). When latencies increase, the fixed number of virtual users will necessarily be sending fewer new requests, leading to a constant or decreasing number of RPS actually being served.

Other load testing tools (Gatling, newer versions of k6) will let you set a target RPS instead.

It's very easy to misinterpret the results of load tests. A common mistake is to run the wrong sort of load test (explained here[1]) and have your loaders automatically back off when latency starts increasing. As the author of the OP touches on in a different context, you're now running a pseudo-limit test instead!

I like the author's emphasis on only proceeding if the server's metrics are clearly within normal ranges.

[1] https://gatling.io/docs/gatling/reference/current/core/injec...

The dodged NI contributions can be passed through via higher comp though - it ends up reducing the total cost of employment.

It's messed up that the UK has such differential tax rates between employment and self employment that this matters at all.

Given that Santa Monica is in the USA, my suspicion is that it's wood frame rather than mass timber as deployed elsewhere, with thick heavily insulated flooring, see e.g. https://www.youtube.com/watch?v=bZGuSC7KNvM

Here's a very interesting video of a Euro mass timber building highlighting some of the differences: https://www.youtube.com/watch?v=bZGuSC7KNvM

This is totally untrue IME of modern mass timber buildings. The sound isolation is excellent.

> 10% of all OK images may be falsely flagged, but what about specific categories of images?

The article doesn't say what % of OK images are falsely flagged, only what percent of flagged images are OK.

Agree that subsets might have different stats, but there's 0 information in the featured article about any of that!

No comment on the policy etc, but I think the presentation of the numbers is a bit misleading.

That 10% is the percent of flagged images which are actually OK. Whether this represents a large fraction of all legal content depends on how much illegal content there is. It would be better if they quoted the false positive rate and false negative rate as a fraction of legal/illegal images respectively.

e.g. if 1/100,000,000 legal images are flagged incorrectly, and 100% of illegal images are flagged correctly, then a corpus of 100,000,000 legal images + 9 illegal images would result in the stats in the headline. That seems like a pretty good system (ignoring any principled objections to the scanning in the first place).

As this page makes very clear, it's actually pretty hard to make a robust webhooks implementation!

What alternatives are there? I've looked at: * Publishing AWS EventBridge events to other accounts. * /events instead of webhooks https://blog.sequin.io/events-not-webhooks/ * ???

F-150s are famously only ever seen on construction sites, never in the Starbucks drivethrough on the way back from the school run.

> But as a foreigner, most of the companies weren't a good fit. Some were too domestic. A lot didn't have any other westerners working there.

This is a great way to think about how important an inclusive company culture is to you - do you want your org to be thought of by non-natives like many HN posters would think of an old school Japanese one? In what ways is it unattractive to non-natives and do you actually value those characteristics?

I like seeing actual numbers about what Good is meant to be for this metrics - it's easy for a devlops advocate to highlight a metrics and push for further investment in it, despite it not being a major weakness of the organisation otherwise.