Catch up on stories from the past week (and beyond) at the Slashdot story archive

 



Forgot your password?
typodupeerror

Comment Re:Never understood how one was expected to contri (Score 1) 118

This is both reason for success of SO and the cause for being the perpetual punching bag. Since there was a high amount of moderation the answers were way better than sorting through forum posts. But people care about their question being answered and don't care that the added volume of duplicates makes the site less searchable for others. And for people answering getting moderated for "good enough" answers is often a stronger deterrent to contribution than fake internet points and altruism is an incentive.

Maintaining a quality knowledge base is at odds with running a community Q/A site. But being the premier knowledge base is the big draw to have people contribute the answers to build the base. And LLMs being able to ingest answers, draws a spotlight on already existing catch-22.

Comment Re:They need better ads. (Score 1) 62

Over a long enough period increased supply usually does follow increased demand,

There's two problematic assumptions here. One is that the facility will be operating at the same capacity long term. Take Microsoft's will they / won't they with Wisconsin data centers or even Foxconn. Businesses can change courses a lot quicker than power plant build outs. So that time horizon is notoriously problematic especially with government incentives in the mix.

Continuing using Wisconsin as a second example, they've have had to delay the decommissioning of the coal plants with expensive results.

That time scale to recoup costs gets a lot longer with an immediate spike in demand that may or may not continue in the long term, but the power plants don't have a choice in whether or not they provide power. Either way, the utility consumers are on the hook for all the risk of increased prices. This is why lower power costs can't be a selling point because they be lowered or they may not, but in the short term prices are going up.

Comment Re: Essentially it is.. an unoriginal con. (Score 1) 47

While it is undoubtedly vendor lock-in, it's not as strong as a lock as virtualization and it's the nature of a complex problem. And even here they aren't pulling the Broadcom of cutting SKUs or tripling costs.

And ultimately, orgs that are using any of the other available tools at scale including FOSS ones are locked in and migration would be painful.

Comment Re:Story checks out. (Score 1) 93

Right, they don't confer sterilizing immunity on an individual scale, which is a large part of the issue with it not being a solely individual choice. Because often the advantages are at a societal scale because of herd immunity, which has lead to eradication (e.g. measles).

So the expectation that a head of policy adequately explains the individual risks and the societal risks is important. As only so many individuals can chose not to get vaccinated before the loss of herd immunity increases everyone's risk. Allowing for informed choices is important, but we shouldn't pretend that it's a solely individual choice.

Comment Re:How is this not against carrier terms of servic (Score 2) 27

The scam farms are doing more than just violating terms of service. And since they are doing illegal activities do tend to move around.

STIR/SHAKEN don't really apply here because this is about having a separate Android device to login into apps like TikTok and Instagram as compared to the scam farms sending bad SMS messages and phone calls where it would apply.

Comment Metaverse (Score 1) 22

Now that the "metaverse" shifting away from being the ruler of the VR headset space, will they rebrand their company again? If so, my vote would be to focus on their core demographics: "Bad AI Memes for Grandmas and Low Child Self Esteem Inc."

Comment Re: Essentially it is.. (Score 3, Informative) 47

It would help if there's a graphics presenting this.

The general premise probably doesn't need a diagram. The idea is that when code is committed that it needs to be "integrated" / compiled and then the built artifacts and executables actually need to be "deployed" somewhere. So besides the storage part of Version Control and the collaboration part (bug tracker, review, merging), then there's the run some stuff like scripts part.

GitHub Actions are that run some stuff part and it's a fairly separate piece of architect (storage is very different than compute). And it's fairly pricy to run that stuff on someone else's servers if you are doing a lot of it.

If you run on your own hardware then you can as well ignore github

So then if someone does have their own hardware and doesn't want to run an entire platform, just wants to compile cheaply on their own kit. Then they could install the runner on their server. However since that runner has to check in to the rest of the platform, there's some minor costs in communicating between the platform and the fleet of runners. And for the vast majority of people, a drop in the bucket compared to user fees. But for that minority of people that are running at scale to have the vendor change pricing models would be very infuriating.

and also self-host your git repo.

There are alternatives, but there's nuance here. An alternative can be to just replace the CI/CD, CircleCI, TravisCI, ConcourseCI, CodeFresh, or even GitLab CI can all hook up to GitHub and then do the builds/deploys. The other alternative is to run the whole version control and collaboration stack, there's GitLab, Gitea, Forgejo, and BitBucket (all that have their own built-in CI/CD systems that for the most part can also be replaced previous list).

But all of that is a lot more work than just running the built-in one that works pretty well and can run on your own hardware to save costs (up until Microsoft tries to stick ya with more of a bill).

Slashdot Top Deals

If you sell diamonds, you cannot expect to have many customers. But a diamond is a diamond even if there are no customers. -- Swami Prabhupada

Working...