I don't claim to have any special skill at AI, but as a 'senior' dev, my strategy is exactly the opposite. I try to be as lazy, dumb and concise as I can bring myself to be with my initial prompt, and then just add more detail for the bits that the AI didn't guess correctly the first time around.
Quite often the AI guesses accurately and you save the time you'd have spent crafting the perfect prompt. Recently, my PM shared a nigh-on incompressible hand-scribbled diagram on Slack (which, in fairness, was more or less a joke). I uploaded it to Gemini with the prompt "WTF does this diagram mean?". Even without a shred of context, it figured out that it was some kind of product feature matrix and produced a perfect three paragraph summary.
I've never really seen the value in the planning phase as you're free to just throw away whatever the AI produces and try again with a different prompt. That said, I don't pay for my tokens at work. Is planning perhaps useful as a way of reducing total token usage?
It's more about the size of the task I try to do, it's quite possible to get opus4.6 to one shot a "good" 30k loc change with the right planing doc. I'm not confident I could get similar results handholding. I also tend to want to know the major decisions and details in such a change up front rather than discovering them post-hoc.
Written like an enthusiast who knows what the filesystem is. The target market of this laptop includes almost no one like you and me. This is for grandmas and teens, and even Gen Z, who is basically 30 now, never learned to use the filesystem, since they grew up using Google Docs and "apps" where all your data is just "in the app."
Apple makes it incredibly difficult to manage usage of the disk, for instance, they only offer you two choices for Photos: All photos locally, or "Optimize disk storage" - i.e. Apple Magic.
iMovie too uses a "Library" which is code for "Apple Magic." If you have a project in progress and want to move it to a USB disk because you need the space on your tiny SSD to download a single game, good luck actually getting that simple task done, because everything including your imported clips and your project files, is in a "Library."
Basic users have no idea how to handle these "libraries," and honestly, neither do I, beyond being able to forcibly move the entire library "bundle" to an external drive, which may as well mean deleting it, since this stuff expects to be left alone in-place.
I mean there are already lots of MacBooks out there with 256GB storage. They haven’t become e-waste because of some mythical inability of users to manage storage (which they already do just fine on phones with even less space).
Biden's surely a poster child for the value of experience and connections in the Presidency. Whatever you think of him (and I would certainly agree that he should never have considered a second term), he was quite successful in furthering his agenda while in office.
Yes, I agree that he used his experience well for many things (and had competent staff he could trust to get things done) but I will say he made a huge mistake continuing to back Israel's actions in Gaza to an extent which I don't think someone too young to remember the Six Days War would have done. I think you could also make a solid argument that earlier in his career he probably would have had more energy to put into getting a few of the close votes in Congress over the line.
I had a similar one from that guy asking me to make open source PRs to some repo of theirs for, err, $25-50/hour. I replied explaining that senior software engineers in the UK aren’t quite as desperately poor as that, and got a canned response saying that they were looking forward to reviewing my PRs :D
SWE salaries in the UK are fairly unsurprising as they tend to be more or less in the range of other professional salaries. If they’re high from a worldwide point of view that’s just because most of us are in London and London is a high cost of living city.
You should be comparing to AMD, at that. There's hope for the upcoming Intel chips - but anything current isn't competitive. Furthermore, M* is good for a specific form-factor, but don't for a second suggest that 4 P-Cores will outdo the 16 hyper threaded cores on a 9950x
The only like for like comparison you can do is comparing an M1 MacBook to one of its Intel Mac contemporaries (which were present day Intel systems at the time).
From my knowledge RCU/epoch/Hazard pointers are useful in data structures and algorithms where raw atomics cannot be used but you still nees lock free or in some cases wait free semantics.
If you can use an atomic then these are overkill and you should just be using an atomic, but many times things that are atomic does not make it lock free, if there's no hardware support the compiler will add a mutex.
I think this is a misreading of the situation. He’s being arrested because of recently uncovered evidence that he committed a crime. We can all form our own opinions on whether or not Andrew committed rape and/or sexual abuse (without too much difficulty, I assume), but this crime looks like it ought to be a lot easier to prosecute.
It's easy for the royal family to say they didn't know he sent an email. It's hard for them to say they didn't know about the girls, due to their security. This way only andrew goes down.
I think this is a very implausible take for two reasons.
First, the King doesn't have that kind of power over the justice system in the UK. He cannot choose when, whether or why Andrew is arrested.
Second, Andrew's arrest is a negative news story for the royal family, not a positive one. Their ideal scenario would have been one where the Andrew/Epstein story gradually dropped out of the news cycle following the stripping of Andrew's royal titles. This arrest, and likely subsequent trial, are just going to keep it in the news for months or years to come.
First, the King approved the arrest because the family has veto power over all actions that effect them. Same happened when they tried to tax the family years ago.
This is a limited hang out negative story. The important thing for the family is that only andrew is implicated not the family as a whole.
The King doesn't have veto powers over arrests of members of the royal family, and he was not notified before the arrest.
In general, members of the Royal Family do pay taxes, so I'm not sure what you're referring to there. (The sovereign and Prince of Wales are exempt by law, but they have paid tax voluntarily since 1993.)
The obvious explanation is that none of the men who abused girls via their relationship with Epstein are being prosecuted in the USA. So why should it be surprising that Andrew is not?
Right, but they're not protected by being members of the British Royal Family. So it doesn't seem necessary to appeal to that factor as part of the explanation for why Andrew has not been prosecuted in the US.
Quite often the AI guesses accurately and you save the time you'd have spent crafting the perfect prompt. Recently, my PM shared a nigh-on incompressible hand-scribbled diagram on Slack (which, in fairness, was more or less a joke). I uploaded it to Gemini with the prompt "WTF does this diagram mean?". Even without a shred of context, it figured out that it was some kind of product feature matrix and produced a perfect three paragraph summary.
I've never really seen the value in the planning phase as you're free to just throw away whatever the AI produces and try again with a different prompt. That said, I don't pay for my tokens at work. Is planning perhaps useful as a way of reducing total token usage?
reply