munk-a

play.shadowsofisildur.com:4500 - if you'd like (it isn't on the old domain anymore) there's also a discord that I'm a bit hesitant to link to where all the discussion is taking place that has a lot of the old folks hanging around and nostalgifying.

The people who played MUDs for the grind will, genuinely, have a lot more fun playing PoE or some other modern adventure game. MUDs were the only option for that sort of adventure experience for a while but there are other options now that simply have more tools at their disposal than text based communication offers.

If a MUD is going to survive today it's going to be driven by those RP focused players where the freedom to describe actions in text allows a flexibility that's unrealistic in graphical games.

It's a method to grade LLM output - as such it's something that will receive focus in correcting for. As soon as people who have a say in where funding is going noticed it as a metric the labs started caring about their performance in it. In the best case the labs are focusing on improving SVG capabilities in general and optimizing Pelican production as part of that initiative - but now that it's a known measure it is no longer reliable.

There is an ongoing project doing this (that launches this weekend I believe) to revive a MUD called Shadows of Isildur. I think a lot of effort has been put into the new changes to the codebase to bring QoL features that a modern audience would expect while keeping the bones of the game the same. I'm not certain how it will fare as, to be honest, MUDs take an inordinate amount of time to really groove into and immerse yourself in - but I am excited to see efforts like this invested on as we older MUDers find ourselves finally reaching the point of having free time again.

A MUD exists as a platform for creative writing - introducing LLMs into one would deprive it of its primary function.

MUDs are still kicking around and if you'd like to play in one you'll find a few solid communities with good engagement that would love to include you. They're in an odd spot though since MUDs existed in two sorts of categories 1. I really want to adventure and text is the only way to do that (these sorts died off to MMOs) and 2. I want to roleplay - those survive but the mechanical portions of those games are becoming more and more irrelevant as new games are constantly supplying better systems and mechanisms than MUDs were able to.

This leaves MUDs in a weird place since things like MUSHes generally have most of the roleplaying tools without as much weird code for attacks and PVP interactions being involved so I've seen those thrive more recently.

Until you need to scale up it's perfectly acceptable to just run postgres on the same instance as the logic that's executing. It's not a great strategy in the long run due to all your eggs being in one basket and the need to configure things like backups manually but it'll save buckets of money compared to going with something prerolled by AWS while the functionality it'd give you wouldn't be noticeable.

cheaper at a geometric rate

Citation needed - my company is paying more than ever for code generation. I have no reason to believe (given anecdotes) that anyone finds themselves in the opposite situation.

It's hilarious to me that one man has so single handedly destroyed the concept of "musk" (the smell) being attractive.

Ah creative writing - the domain where AI is failing to actually generate any economic value while absolutely devastating the ways in which people have traditionally found artistic expression.

Yeah - our multi TB scan from two decades ago was a dolphin brain image which is a fair bit larger than mouse - and it was a stained sample that ended up being used as our demo image frequently because the contrast dyes set well and resulted in a very pretty visual overall.

I didn't meant to say that PBs of image data is common place - we had no image that approached that size - but people outside the domain of microscope scan results might be unfamiliar with just how chonky these image files would get traditionally.

I used to work with images representing scans of brain tissue - for a full brain visualization at one horizontal slice terabytes was a common measure and the resolution of those images wasn't even particularly detailed - all the full resolution stuff was taken of tiny sub-sections of interest. This was also two decades ago - so I'm sure they've upped their game.

And, of course, Grok: Sir-not-appearing-in-this-listest

It's also important to note that several PE firms have signed contracts with model trainers to specifically use their model in their owned companies. I know Anthropic signed a deal worth hundreds of millions to acquire users just a few months ago.

Most of the folks here on HN are dealing with customer feedback in systems automation in one form or another - it's pretty unavoidable in this age of LLM trendiness. The customers of healthcare (in both private and publicly funded systems) are the patients. So while the term might not be super natural it's an understandable one to use.

Yeah, I didn't even want to lean out that far but I'm sure average handle time is incorporated in some manner.

But, just purely semantically, the statement Kaiser gave in response was worded in a precise smug corporate America style to dodge the main concern raised. I think it's important to call out weasel words.

"Nurses fear that having long calls can lead to bad performance reviews"

A company spokesperson said, "Kaiser Permanente does not use Average Handle Time to assess agent performance"

So uh, average time wasn't raised as a concern, calls beyond a certain threshold was. I wish this semantic discrepancy was better highlighted in the article.

Just to comprehend this a bit better - it sounds like the FAA had stripped Boeing of the ability to self-recertify and actually sent inspectors for the most recent certifications. After several successful certifications and what would appear, to the inspectors, to be real process improvements, they're now re-granting Boeing the ability to self-recertify when self-recertification is allowed?

This is well outside my knowledge domain so I'm not trying to make any statements on whether this was correct, but rather to better comprehend the change.

The fact that body cameras can be turned off is insane to me. Most office workers need to suffer the same "constantly surveilled during work" and the stakes of their mistakes is so much drastically less than law enforcement officers.

Releasing full unredacted body camera footage continuously is a bad idea - but prosecutors and defense attorneys have been trusted with equivalently sensitive information in the past. I don't know how we lost the battle on this front and allowed them to ever be turned off.

Modern car thefts are just a side effect of some terrible design decisions though. If you buy a car with a physical key required to unlock it the likelihood of it being stolen is severely reduced - it's not zero, of course, window smashing and hotwiring are still issues. But the fact that a continuously transmitted signal can be used to unlock and start a car is such a dumb idea for such a small gain in convenience.

With well considered engineering it doesn't even need to be tap water. If you have a closed loop thermal conductor that interacts with the components themselves you can then use really trashy contaminated water that just needs to be clean enough not to actively erode the heat transfer mechanism. We have setups like this all the time that use condensed air via cooling towers or salt water immersed heat sinks to discharge energy - it's more expensive than tap water but it isn't technically complex. So if it ever becomes unpalatable (likely due to politics) to use tap water there are some readily available alternatives.

The big win of being in space is just a worse alternative to using an intermediary heat transfer medium.

A fair amount of ML/AI innovations came out of the market in general. Neural networks are a useful tool to solve a variety of problems... LLMs specifically were a more recent interesting market to develop but I've yet to see anything that could give a market player a real competitive advantage. It feels like we just invented a new hammer and now that we know how to build it it isn't that hard to build one yourself. The all purpose hammers are, of course, unreasonable to build - but those don't seem to be that useful. I don't really need Claude to be able to generate sonnets when I'm programming so I think specialization is the place we'll see genuine markets form.

LLM training doesn't carry the same NIH risks that normal internal software bloat does. They are relatively simple to setup training for and analysis of accuracy/recall can be automated.

This leaves the price differential between a private third party and an internal initiative as barely more than the cost to train the model[1] - perhaps that's where we'll end up, a centrally trained model will represent an economy of scale that can leverage that difference into a margin it can profit off of but your business being purely profit driven by that training expenditure seems like a ridiculously thin margin.

So where does that leave the AI companies? If their LLMs are off the shelf-once built products they have a strong advantage for casual low usage but enterprise customers will have a huge cost incentive to roll their own - if the LLMs require continuous retraining and the frontier keeps moving then enterprise customers will find a packaged service more attractive and likely continue to subscribe for more accuracy but casual low usage will likely shift towards "good enough" models. It seems inevitable that they'll lose half the market and it seems difficult to discern their long term profitability[2].

1. Costs can, I think, reasonably be reduced to hardware depreciation and energy - if trends continue with cloud resource availability (it's possible this won't be the case if large compute providers start pulling resources offline to build a moat but I think they'd likely prefer the reliable compute income over model income which has several other competitive weaknesses). Hardware depreciation would normally be pretty negligible and equal across different training entities, right now we have a chip shortage but given the demand that can't last too long so I'd consider hardware to be fungible - and energy is entirely fungible - they're both hard to moat.

2. Outside of AGI, who knows if AGI will be or what even counts for it at this point - but I think if AGI isn't a doomsday scenario we fall back to one of the two above scenarios - either the frontier is ever moving and they can retain enterprise customers or the frontier seizes up and everyone can just use an off the shelf offering. In either scenario they don't have a lot of moat to deal with for their products unless they can restrict compute which is why Alphabet, AWS and MSFT are the only players I could see realistically coming out of this as an AI vendor winner and I'm not even certain if it'd be a good idea for them if it'd hamstring their cloud profitability.

It's true, if you're spending that much money you must be worth a ton! Just look at SpaceX!

Elon initially sold xAI as having a spicy mode and being politically incorrect.

It was only deemed a bug when it became a liability - you can't simply rewrite history and expect it to go unnoticed.

Even if you personally have no qualms about Elon Musk his PR is a mess and introduces a lot of risk for long term company viability and funding that competitors just don't have.

This ruling is very early in the process and doesn't prevent anything - but given the more likely bad path outcome of this case for OpenAI it wouldn't disallow their usage of OpenAI but instead prevent them from shutting down competitors that claim to offer an open AI.

An excellent example of something that clearly wouldn't be distinctive since all AI related endeavors bill themselves as unicorns.

Even the US does change policies every once in a while. This is a case where Open Systems was last challenged under the old rule set when it was acceptably descriptive. Since that point standards have changed and, presumably, Open Systems might now be exposed to a similar mark challenge.

Sugar in moderation is a fine thing the issue the US has is more focused on how pervasive sweeteners have become in what looks like savory food. A ban would be a very silly thing but at some point America needs an FDA with teeth to actually crack down on labeling requirements.