capableweb

Depends on the workload really. Some workflows involve copying large amount of stuff from disk to RAM and then over to VRAM on the GPU, some workflows are really inefficient while others require basically no VRAM at all.

For example, for fast 3D renders of small scenes you just care about processing speed, large scenes need good memory bandwidth too, while with LLM training you'd care more about available VRAM, and so on.

But, just to throw out a number out there, you could probably get away with spending as little as 300-400 USD for just a motherboard, PSU, CPU, cooler and RAM, assuming you don't want things like storage, cases and whatever accessories. Basically use it as a server.

Would be interesting to see if it actually gets better performance. On paper, it should, but famously when you control the entire stack it's easier to optimize so software and hardware works well together, and that's exactly what Valve has with the Steam Deck.

Personally not blocked at all as those two are things I use in a professional setting related to graphics work, and then I'm forced to use Windows regardless of what I think.

Sure, but there are also legitimate uses for things like that, otherwise people wouldn't want them. How you build screen-recorders for example without "grabbing pixels that aren't yours"?

Clearly, something is missing here. Why not a permission system like browsers do? A little popup that asks "X is requesting access to your full screen" rather than this "convenience VS security" black and white view most people seem to take a stance at?

Additionally, Naughty Dog famously been pushing/punishing gaming systems for a long time, even games like Crash Bandicoot (1996 - PlayStation 1) and Jak and Dexter (2001 - PlayStation 2) really went up against the limitations and lots of hacks were incorporated to get the games to run well.

Uncharted and The Last of Us is just continue what is tradition at this point :)

Maybe it's worth responding directly to the criticism from Zig rather than calling what they are saying a "rant".

Actually, since you don't address anything in your comment, one could call that one a "rant" as well.

> but short clips taken from movies have been on youtube for more than a century.

AFAIK, it seems like entire movies are being split into 40+ parts and uploaded to TikTok, no just "short clips taken from movies". That doesn't really seem like "fair use" to me.

> Norwegian Transport Minister Jon-Ivar Nygard said in the statement:

> "Electric cars give us greener transport, but they also have a clear intermodal competition with public transport in urban areas. We must make it more attractive to travel by public transport, cycle, and walk."

I don't find that statement anywhere though, but that's from the article. I think that means they're working on doing something about it, but haven't reached any proposals yet.

Me too!

Hello! That's probably a sign I need to take a break from writing too many HN comments per day, thanks :)

Well, I'd argue that it's just two different names for similar concepts, but applied at different levels. WAL is a low-level implementation detail, usually for durability, while Event Source is a architecture applied to solve business problems.

A WAL would usually disappear or truncate it's length after a while, and you'd only rerun things from it if you absolutely have two. Changes in business requirements shouldn't require you to do anything with a WAL.

In contrast, Event sourcing log would be kept indefinitely, so when business requirements change, you could (if you want to, not required) re-run N previous events so you can apply new changes to old data in your data storage.

But, if you really want to, it's basically the same, but in the end, applied differently :)

Building a storage service like these today and not having "append" would be very silly indeed. I guess S3 is kind of excused since it's so old by now. Although I haven't read anything about them adding it, so maybe less excused...

Or, a more hopeful view, is that "ad targeting privacy" makes "privacy" be a topic at all in society, and that could be leveraged to acquire more privacy in more areas, as people start to realize what it actually means.

Problem is that "evil marketing companies" is a lot easier for people to believe in, compared to "evil government", especially in places like in the Nordics where people generally have a much more favorably view of their governments than in other countries (for better or worse).

That's a pretty good point actually.

First solution: Yes, finding the PC CD for this game took me like 20 minutes, both new/old in even my country for something like 20 EUR. If the author wanted to, I'm sure they could have found it equally easy.

Second solution: Not really, when you do CD releases you have a "gold master" which is the final burned CD that the other CDs are copied from. That's the one you use as the official one, as that's the one most people on PC actually have.

If the game had multiple regions (something PC games usually don't), you have multiple executables. But this happens mostly for console games, but again wouldn't be difficult to acquire a specific one for one of the ~4/5 regions there is.

Re third problem, you're probably right about this. But I'd still argue that verifying information after a finding would add more credence to the story itself and cement it better, it wouldn't take anything away from the story (unless it proved that there isn't actually any story).

> Append workloads are common in distributed systems.

Bringing back the topic to what the parent was saying; since S3 is a pretty common system, and a distributed system at that, are you saying that S3 does support appending data? AFAIK, S3 never supported any append operations.

> The criteria for trademarks is not defamation, but confusion.

I'm well aware of that, I'm mentioning two potential issues, one being the "confusion" part and the other being the "defamation" part.

Last time gcsfuse was on HN (https://news.ycombinator.com/item?id=35784889), the author of rclone was in the comments:

> From reading the docs, it looks very similar to `rclone mount` with `--vfs-cache-mode off` (the default). The limitations are almost identical.

> However rclone has `--vfs-cache-mode writes` which caches file writes to disk first to allow overwriting in the middle of a file and `--vfs-cache-mode full` to cache all objects on a LRU basis. They both make the file system a whole lot more POSIX compatible and most applications will run using `--vfs-cache-mode writes` unlike `--vfs-cache-mode off`.

https://news.ycombinator.com/item?id=35788919

Seems rclone would be an even better option than Google's own tool.

I mean, duh? Anything that gets breached and leaked will obviously end up in various lists, and has always been doing so.

> In November 2022, the password manager service LastPass disclosed a breach in which hackers stole password vaults containing both encrypted and plaintext data for more than 25 million users. Since then, a steady trickle of six-figure cryptocurrency heists targeting security-conscious people throughout the tech industry has led some security experts to conclude that crooks likely have succeeded at cracking open some of the stolen LastPass vaults.

> longtime cryptocurrency investors, and security-minded individuals

I'm not sure how "security-minded" you are if you, months after a breach of your password manager, still haven't changed all the involved passwords and keys, especially those involving things worth a lot of money!

> documentation for human consumption will be an afterthought.

The documentation just changed a layer upwards in the abstraction ladder. Instead of docs for the code that runs on the machines, we'll write docs for the programs we've built that can generate the code that runs on the machines.

You need to either be or represent the copyright holder. I'm 99% sure most warez groups have no interest in exposing themselves to any legal systems when it comes to what they do.

Why is this hex dump being used as a verification rather than just showing the two (supposedly the same) hashes of the entire executable themselves? Should be a quick 20 minute job for whoever call themselves "journalists" today.

I suppose it's possible for that particular string to get into the binary by some other means. Maybe failed anti-circumvention that checks for known strings in its own directory or something similar?

A mantra I usually follow is "If you're putting something under 'utility'/'utilities'/similar names, you're cheating on your design and you're only putting it there because you can't figure out a better place, think longer/harder.". Applies equally for testing code :)

I'm guessing the issue is that it's obviously "inspired" by "Terraform" which shorthand is TF.

My second guess is that using "Open" together with something that implies "Terraform" is also a problem, as Hashicorp might sue them for defamation, as that name would imply that Terraform is not "Open". I'm not personally against that, I wouldn't consider Terraform "Open" anymore, but I'm not sure a court would see it the same way.

I'm fairly sure this is because of the longer context window they've probably enabled for GPT4 on ChatGPT, I'm seeing the same thing.

When you use the longest context window, it tends to place the most focus on the first and last message, for whatever reason. So if you instead of continuing the conversation, you start it from the beginning but with the additional information gained from the previous one in the initial message, it does a lot better.

> our predecessors spent so much time in getting everything 100% correct from the ground up

What predecessors are you talking about here? The internet is barely held together with duct tape and servers running only because of minute by minute injection of WD50 straight into the fan bearings. There is so much cruft everywhere that is being relied on daily in production systems that most people lost count by now on how much horrible code and setups they've seen in the life.

Tutoring makes sense, arithmetics doesn't. Instead, what you want, is to include something like "If you need to evaluate formulas, embed them in a Markdown code block with the language set to 'math' which will be executed for you, and the results will be provided as the next message" to the system prompt, and write code around it to handle the execution.

Simplified example:

> system-prompt message

> user message: teach me linear algebra

> assistant: Sure, here's how [...] ```math 1 + 1```

> user message (but generated by application): ```1 + 1 = 2```

> assistant: and since 1 + 1 = 2, [...]

LLMs are great at generating text, but not at math, so make something else do the math for the LLM and then it can focus on what it does best.

> the response was it never goes down.

Seems those folks maybe have never used GitHub in any larger capacity. They've been having issues constantly for a long time now.

September - 7 incidents

August - 17 incidents

July - 11 incidents

June - 13 incidents

And so on. Last time they had less than 10 "incidents" was December 2022.

I'd like to believe I have some clue, yes.

It feels like you allude that I don't. Is there something specific you'd like to mention here, or am I supposed to guess?

The "TF" part could be anything! :)

Open Technology Foundation is one suggestion.

Backronyms are a thing people.

If a company isn't ready to compete against someone/something that is using the code the company has written and published as FOSS, then yes, please do not use FOSS licenses for your "Open Source" product.

One would think that the companies thought this through before publishing FOSS code, but seemingly there is a lot that didn't do that.