Too

All editors do this already.

There are several topics like this on the front page every day, since many months back now, and the comments are all the same. It’s enough.

> 1/100th the cost with better reliability and performance

Cheaper yes. More reliable? Absolutely not. Not with today’s models at least.

”Attractions along the route” must have been a standard feature on satnavs since the very first standalone TomTom sucked onto your windscreen. Newer cars having Google Maps built in make this even richer.

If by ”nowadays” you mean the past 30 years. Slow network drives making Explorer go completely unresponsive has been a thing since Windows 95.

I’m more surprised to hear that bug still hasn’t been fixed. Luckily I don’t use Windows myself since many years ago.

git rebase -i lets you reorder commits easily, as long as they don’t have conflicts. If they do, you’re in for a rough time. I struggle to see how a vcs could help with that, even if I’d be happy to be proven wrong.

One cool tip to help with conflicts is the revert trick. If you have a conflict that you need resolved earlier in your commit chain, you can commit a cleanup that hides the conflict, revert it instantly and reorder the commits with interactive rebase to insert the revert first. It’s a bit hard to explain without an example, once you’ve tried it you will understand.

One exception to this is batch scripts and other cli tools with a clear start and end, like an installer, rsync, curl, dd, etc. Setting up metrics here is way overkill and the user may still be interested in the progress. Easiest made available through logs. Curses UI could be a nice middle ground but also very often overkill.

When inference requires maxing out the memory of a gpu, where are you planning to keep this cache? Unless there is a way to compress the context into a more manageable snapshot, the cloud provider surely won’t keep a gpu idling just for holding a conversation in memory.

So instead of ’const tubeStations = useDatasource()’ we now type ’<DataSource id=”tubeStations” />’ and anything beyond this basic example is left as exercise for the reader.

Doesn’t seem like a big improvement? In fact the opposite. Sure, React’s stateful hooks mixing in a semi-declarative mode of programming in otherwise imperative code isn’t exactly intuitive, but the flexibility and familiarity you get from having full JavaScript at your fingertips is worth everything. Hands up anyone who want to learn how to make a for-loop with filter+map in xml syntax. Didn’t think so.

If you want to fix react what you should do instead is keep jsx and give stateful hooks a dedicated section that can’t be mixed up in rendering code.

Can someone try this in a virtual machine? I’d imagine syscalls there to be more expensive.

Did the same many years ago and felt equally liberated.

The thing i found about most of my notes was that they quickly got irrelevant, "PHP and MySQL best practices, on Windows..."? First, the notes that were referred to often, they quickly got upgraded to a script or stuck to my first brain so that note wasn't needed any more. Second, finding it through Google was easier than searching through my personal notes, and would often yield some new insight I wasn't aware of. Learning how to read official documentation rather than blog posts was a large part of this journey as well, now I trust that no voodoo tricks are needed to get things flying if you just follow the standard process. Being savvy enough to debug and lookup anything helps. Third, the tools, or their best practices, got superseded by others.

Articles that might be interesting to read (or re-read) one rainy day. Forget about it. Stop the wishful thinking. Every week there are new ones, more up-to-date and equally insightful. Your future self will also have other interests. This was by far the most liberating.

Nowadays, the few things I keep are topics that don't change, that I know will be difficult to remember between the rare occasion I need it (annual domain renewal, spare part number for oil filter). One final thing i still like to hoard is material from experts about recurring religious topics. So that next time someone at work starts arguing about monolith vs microservices again, I can just link to authoritative source instead of having to reiterate the same old arguments and prove my expertise in the area again.

That only works if you let the LLM do the interpretation of the MCP descriptions, in the case of TFA the idea was to use MCP without LLM, which is essentially same as any old API.

In embedded you are typically stuck on some ancient proprietary compiler and can't take advantage of the latest C versions. Even less so if you need safety standards like MISRA.

That of course doesn't help you with the switch away from C. The question is why they keep updating the language. The only ones with valid reasons to not upgrade to some more sane language can't take advantage of the new features.

> I upgraded from 10 to 11 on existing hardware.

Good for you. There is plenty of hardware out there without TPM 2.0, that is not allowed to upgrade, even if they in every other aspect are more than capable enough.

Do you have a less hypothetical example to share?

Just a basic prompt that makes use of this server and how it responds. Or a simple agent conversation that continues successfully beyond 5 roundtrips.

Yeah. I can't get the amount of hate seen here either. It's super natural to read and has very few concepts to learn. Just resources, lists, sets, dictionaries, variables, ternary operator, function calls and templates. Paired with a LSP and syntax highlighting it gets even easier to see the boundaries between code and variables. The two different ways to write dicts are a bit strange and foreach has some gotchas but honestly not a big deal. Especially not when put in comparison to YAML which has nothing even close and even more invisible footguns. Seeing comments like "it does not have if-statements". Duh, it's not imperative, for many good reasons. Not understanding that yeah it's clearly a skills issue.

Replace docker with podman and you can go the other way around, run a pod manifest locally.

Interested to hear more about this. Care to elaborate more? Maybe there are already articles written about this?

You could have sane defaults for something running locally. Take Jenkins as example, all you do is start it up and give your agents an SSH key. That's it, you get a distributed execution environment. Not saying Jenkins does what K8s does, but from an operational standpoint, it shouldn't have to be much different.

I invite anyone to try reading https://kubernetes.io/docs/setup/production-environment/. It starts with "there are many options". Inside every option there are even more options and considerations, half of which nobody cares about. Why would i ever care which container runtime or network stack that is used. It's unnecessary complexity. Those are edge cases, give something sane out of the box.

Say what? Protobuf may be type safe but it is far from a configuration language, can you even refer to variables or do loops. Unless you are thinking about serializing it with builders in code, which takes away the benefit of declarative configuration. It's also not readable without tooling. As parent said, this really sounds like refusal to learn something new.

It allows the exact same problem. version: 1.30 vs version: "1.30".

> "Kubernetes isn't opinionated enough"

yes please, and then later...

> "Allow etcd swap-out"

I don't have any strong opinions about etcd, but man... can we please just have one solution, neatly packaged and easy to deploy.

When your documentation is just a list of abstract interfaces, conditions or "please refer to your distribution", no wonder that nobody wants to maintain a cluster on-prem.

Ergonomics of using the right data structures and algorithms can also play a big role. In C, everything beyond a basic array is too much hassle.

How can one know that's not coming from the pre-trained data. The paper is trying to evaluate whether the LLM has general problem solving ability.

Is it normal to have human operators pointing the camera around like that? It almost look like they expected something to happen.

Yeah. It doesn’t take much imagination to figure out that a Coke Zero looks identical to regular Coca Cola, yet they have vastly different calorie content, to put it mildly.

Sauces loaded with butter, sugar or other goodies will of course be the same story.

Index of files stored in git pointing to a remote storage. That sounds exactly like git LFS. Is there any significant difference? In particular in terms of backups.

Care to explain more?

> combination of the manufacturer, make and serial number.

> patent that involves defining that as a unique identifier for aircraft.

Now i got mighty curious what makes this novel enough to be a patent.

The million dollar question is how far can one get on this trick. Maybe this is exactly how our own brains operate? If not, what fundamental building blocks are missing to get there.