But it's the reality of engineering. If reality is unacceptable, that's not reality's problem.
But the problem is, the sales world has its own reality. The reality there is that "we don't know when" really is unacceptable, and "unacceptable" takes the form of lost sales and lost money.
So we have these two realities that do not fit well together. How do we make them fit? In almost every company I've been in, the answer is, badly.
The only way estimates can be real is if the company has done enough things that are like the work in question. Then you can make realistic (rough) estimates of unknown work. But even then, if you assign work that we know how to do to a team that doesn't know how to do it, your estimates are bogus.
I'm not aware of anything. I have to read between the lines when experts (some of them university professors in agriculture) talk to figure out the above.
> Because the semantics of numbers and dates matters.
Type semantics is only a small part of what is needed for systems and humans to know how to adequately work with and display the data. All of that information, including the type but so much more, can be supplied in established ways (more graphs!) without having to sprinkle XSD types on your values.
For example, say you have a value that is expressed by a number that for whatever good reason can only be between 1 and <value from elsewhere in the graph> and only in 0.1 increments. Knowing that it is a number and being able to do math on it is not that useful when 99% of math operations will yield an invalid value; you need that metadata, and once you have it you also have the type.
Seattle public library is also an archive as well as a provider of many beautiful and free third spaces. The downtown library is very cool. I bet there’s stuff in the stacks there that is not digitized anywhere.
> the argument is they get you out of the copyright infringement caused by copying the software from disk to memory
This is not copyright infringement in the USA:
> …it is not an infringement for the owner of a copy of a computer program to make or authorize the making of another copy or adaptation of that computer program provided… that such a new copy or adaptation is created as an essential step in the utilization of the computer program in conjunction with a machine and that it is used in no other manner
I have to say I work for John Deere to discuss this more, even though I don't speak for the company.
A DOT (I'm not sure which DOT) just did a press release on how they used John Deere guidance on a snow plow which allowed them to clear the road in a blizzard so an ambulance could get to the hospital (I was surprised they can get enough of a GPS signal, but apparently they did). Auto steer allows someone to drive a plow when you can't see the pavement/lines without first having to memorize the road by the posts/trees on the side of the road.
However there is a big difference between Deere auto steer and Tesla FSD: safety. Tesla has sensors to see if someone/something is in the way and algorithms to go around - critics claim they don't work well, but they work infinity times better than the complete lack of any of those sensors/algorithms in Deere's system. If you are using the Deere system it can hold a lane to within a couple cm - but you have to look out the window constantly because it will just drive right into anything in the way. This is good enough for farming (nobody/nothing is going to be in front of the tractor anyway), or the DOT (they can't see the road at all, but they still have trained operators ready to hit the brake) - but Tesla is going after the "you can take a nap" market.
I wouldn't be surprised if Deere has more miles of self driving than Tesla and Waymo combined, and a better safety record. However this is only because Deere's system is used in situations where the odds are against there being anything to harm in the first place, while Tesla/Waymo are trying for the much harder open road with who knows what in the way.
Now Deere is working on the full autonomous solutions, I'm not sure what the status is (I think some are out there for use in very limited situations). I'm not allowed to say anything more about these plans (I know some is public but I'm not sure what)
It's a next-word-prediction-machine, not a calculator. It's not aware of the passage of time, or how long things take, and doesn't reason about anything. It's just very good at putting words together in combinations that look like answers to your inputs.
That's really useful for some tasks, like regurgitating code to perform a specific function, but it's basically useless for jobs like estimating schedules.
Choose 2. For example a large feature set can be made quickly, but it will be of poor quality.
Note that cost is somewhat orthogonal, throwing money at a problem does not necessarily improve the tradeoff, indeed sometimes it can make things worse.
If you have complex tasks and you have more than one person put in time to do a proper estimate, yes, you should sync up and see if you have different opinions or unclear issues.
Always send "pragma foreign_keys=on" first thing after opening the db.
Some of the types sloppiness can be worked around by declaring tables to be STRICT. You can also add CHECK constraints that a column value is consistent with the underlying representation of the type -- for instance, if you're storing ip addresses in a column of type BLOB, you can add a CHECK that the blob is either 4 or 16 bytes.
Fair point on HBM, but in a production system the orchestration layer is often the actual bottleneck. I've found that keeping the GPU fed requires a level of concurrency and stability that is hard to tune in Python. Rust is useful here not for the inference itself, but for ensuring the request pipeline doesn't choke while the GPU is waiting for data.
I may have the answer. This website nand2mario and this article was here on HN a while ago. Maybe a month or two months ago. So this is a dupe posting. So maybe then you were busy and didn't read it fully but your mind caught it somehow from HN or coworkers/friends and now when you are relaxed you can register the stuff nicely. Just type 80386 in the search and you'll get that previous post.
Note that ~3-6ns is on modern desktop CPUs where extra few kB matter less. On microcontrollers it will be larger in absolute terms but I would expect the relative difference to also be moderate.
Why I built it: With the rise of "Vibe Coding" and AI tools, shipping new products has never been faster. But that also means "abandonware" is piling up. As a user myself, before I pay for a subscription (or even consider buying a micro-SaaS), I constantly ask myself:
"Is there still an active developer behind this?"
"Is this project alive, or was it abandoned 3 months ago?"
I built GitPulse to answer that. It’s a verified signal for users and buyers to see if a project is actively maintained or stable without exposing the source code.
Features I added for fun & utility:
Gamification: "Shipping Streaks" to keep myself motivated.
Social Sharing: To show off progress publicly or prove activity to clients/buyers.
Smart Context: Distinguishing between "High Activity" and "Stable" (so finished projects don't look dead).
Roadmap: I’m currently working on deeper metrics like Build Stability (CI/CD status) and Package Health, so it’s not just about quantity, but quality.
I don’t think that has anything to do with not being able to buy a house or have a child. TCG cards are the perfect mixture of consumerism and gambling, and Gen Z has been submerged in both for the entirety of their lives
The alternative is just not having FDE on by default, it really isn't "require utterly clueless non-technical users to explicitly opt-in for backups to avoid losing all their data when they forget their password".
> Have we got to the point where we need an article telling us that slighting people doesn't help their motivation?
American culture is unfortunately permeated with examples, and habits, and expectations around punishing the behaviors you'd want to see. I see subtle things like that all the time. So while I doubt anyone who stopped to actually think about the concrete implications of their behavior, more specifically their unconscious habits; wouldn't be able to describe how insulting people, or really, how discouraging people is likely to have a negative outcome. The catch being, most people don't stop to consider anything. Thoes who do, are exceptionally rare.
As an example, someone posted a comment providing context, and encouraging people to be curious and grow their skill set with techniques that will help them with dogs, (and yes, these do translate to humans as well.)
Which invited a negative comment from you attacking people who aren't perfect every single moment of every single day, who might benefit from a reminder that how they treat others matters. Also indirectly attacking the person you replied to.
(See what I mean about the culture of punishing the behavior, you want to see? Or did you intend to discourage curiosity?)
> Perhaps the answer is yes when we also compare a worker's motivation to a dog's motivation seemingly without irony.
You can train a human using the exact same skills you use to train a dog. Just because humans are also, in addition to those able to do a lot more, and learn in an astronomically larger set of ways, doesn't exclude the techniques that work best with dogs. You forget this at your own peril. I.e. if the way you behave wouldn't encourage the behavior you want from a dog well, it sure as hell wont encourage the behavior from a human. All humans, including you, are not that special, get over yourself. So to speak
But the problem is, the sales world has its own reality. The reality there is that "we don't know when" really is unacceptable, and "unacceptable" takes the form of lost sales and lost money.
So we have these two realities that do not fit well together. How do we make them fit? In almost every company I've been in, the answer is, badly.
The only way estimates can be real is if the company has done enough things that are like the work in question. Then you can make realistic (rough) estimates of unknown work. But even then, if you assign work that we know how to do to a team that doesn't know how to do it, your estimates are bogus.