solid_fuel

Only in the context of something like RAG - there, you can attach provenance information to data included in the prompt.

With training data, however, no such annotation is possible.

I would rather keep things on postgres but everything wanting its own postgres container is really annoying. Postgres already has support for multiple databases, users, and access controls. I would personally prefer to just set up one postgres instance with proper backups and have everything run against that, unfortunately some self-hosting projects have fairly poor support for that.

Well yes, but that's the point entirely - you can easily put a binary into a docker container, or a jail, or on a VM. But when something is shipped as 5 different docker containers it is MUCH harder to extract it into something you can run without docker or in a different configuration (like sharing a postgres instance between multiple applications).

One distribution method is flexible and can be deployed however you want, the other prescribes solutions and makes any customization difficult.

You might not like hearing it, but it is true.

Code has never been the bottleneck for a skilled engineer, and if you value high quality then you need to spend a lot of time reading over the slop LLMs produce to turn it into something hardened and reliable. So, for a skilled engineer, LLMs have a very marginal benefit in terms of delivery speed.

This is backed up by LLM productivity studies.

> Here the data model is a direct copy, down to the terminology.

That’s not surprising to me - they designed and built it using their LLMs, of course it’s duplicating prior work. That’s essentially the core of what LLMs do, after all.

What a silly response. The parent post is factual, your GP post is not.

There, now we have all brought the same amount of rigor to this discussion.

I’ll also point out that everyone sees Russia struggling in this war that they’re “definitely winning”. The baseless assertions may have been intimidating at the start of the war but we’re years into it now and Ukraine is still standing.

> Investigators also sought permission to use an invasive surveillance tool called a cell-site simulator against journalist Georgia Fort and a Minnesota state Senate candidate.

There, I fixed it for you. Now you can freely read and engage with the content of the article.

> Autonomous AI Agents are failing in enterprise environments because standard Vector Databases (Pinecone, Milvus) act as static filing cabinets. They rely on discrete Euclidean metrics like Cosine Similarity, which blindly retrieve contradictory information (hallucinations)

Hallucinations are a result of the next token selection process, not simply caused by bad RAG. In effect, all LLM generated content is a hallucination, sometimes the hallucinations are useful.

Eliminating “bad” data from your RAG database is not going to fix hallucinations.

Dreams are one thing. People here are tired of fantasies and wishful thinking. Looking at the problem here - Adobe treating customers like crap and exploiting them - most of us realize that the solution is simply regulation. Leveraging the power of society and government to force the company to behave properly.

Gesturing at LLMs and going “this and the free market will fix it” is laughably naive, especially in 2026. That is why people are reacting harshly. This is a problem that was created by the free market, it’s not magically going to be fixed by some new product.

The New York times publishing an article you don’t like is not “communist infiltration”. In this country, we have a right to free speech, including free press.

Somehow I am not surprised to see you repeating the red scare here. This attitude is anti-intellectual and destructive.

> We need a more invective word than 'pseudo-intellectual.'

Here's one: 'AI-Bro'

Criticizing a subset is not the same as criticizing the whole set. You know this. Pretending otherwise is a waste of everyone’s time.

All this requires is basic reading comprehension and a willingness to not read into statements things that weren’t stated.

Yeah LLMs do have a lot of similarities with markov chains.

You may wish to read the thread more carefully, because you are mistaken. If after carefully rereading the thread you still believe you are correct, please explain exactly what part of this sentence:

> “And Zuck’s sudden interest in MMA/boxing/Jiu Jitsu as part of his “alpha male” rebranding”

equates general interest in fitness with being an “alpha male”. Because to me it is pretty clearly referring to the interest in combat sports, not fitness.

>> And Zuck’s sudden interest in MMA/boxing/Jiu Jitsu as part of his “alpha male” rebranding.

> Huh, taking care of your body, being interested in a sport, and not being a robot is now labelled as an alpha male?

You're doing a lot of projecting here.

Says who?

Why shouldn't our civilization be defined by how efficiently we use the energy we gather? Or the lifespan and health of our citizens?

Claiming that the sum of a civilization is how much energy it controls is just a meme from the 50's. It's a shitty, short-sighted metric. One that would consider building a nuclear reactor and dumping the power into a bank of resistors as contributing "more progress" than growing abundant food, or protecting the ecosystem.

It's our civilization, we can define progress however we want.

What's that? The mechanical parrot can be tricked into repeating things that aren't true? It's almost like these are statistical language models and not literal Deus Ex Machina.

> If you want to know why big, progressive cities have the most unaffordable housing, look no further than this law.

And how exactly does the causality for this work, in your head? Rent is already high in the big, highly desirable cities, before this law.

Is this some new form of retro-causality? A law passed today drove rent up 5 years ago?

Seriously. They already get caught uploading everyone’s private credentials once before, one would have to be a particularly gullible rube to trust grok again. Especially with musk in charge.

> Snarky comments deserve snarky replies so there you have it.

You started the snark. Shallow dismissals do not merit detailed replies.

> You won’t have to take the Trump administrations word you’ll just have to read the agreement and accept the facts of it even if it doesn’t fit your preconceived political narrative.

Here's an idea: Why don't you read the agreement and if it is actually a significant change then you can come back with some facts instead of just blindly trusting the administration just because it makes you feel good.

Until some actual evidence is presented, none of us are required to accept the word of people who have demonstrated repeatedly that they are allergic to the truth.

> so why are they seemingly hiding a small number like this?

Reflexive lies are basically the policy of this administration and party. They have demonstrated no interest in telling the truth about anything. The proper question is: Why would they start telling the truth now?

Their base has already demonstrated that there is no red line. There is no lie they could tell that would suddenly disillusion the MAGA-types.

> False accusations of racism (and similar moral slander) is the first step in dehumanizing an outgroup.

Accurately describing the racism that motivates the politics and actions of a group is the first step in addressing the way they dehumanize their opponents.

They put an alcoholic fox news host in charge of the military and he fired or otherwise forced out a ton of experienced leaders in the name of "the warrior ethos" and "high testosterone soldiers".

In a completely unrelated development the "worlds best logistics organization" is now having trouble with logistics.

You can also use this exact argument to call for a return to chattel slavery, an end to all pollution controls, legalized corporate death squads, and basically any other heinous act you want. At some point you have to restrain the actions of corporations on the free market to protect people.

If you're really looking for causes, the largest one is price optimization.

But, I think one missing piece between architects and the general public is how architecture students are taught to design (and thus, see) buildings first through the lens of massing [0]. When you compose a building primarily by shaping the different masses, you lose sight of the ornamentation and complex decoration that sets old buildings apart. Then, you add air conditioning and modern building systems and the need for high ceilings and windows with airflow also disappears.

What you are left with is a sealed box. Essentially, architecture has ceded too much design space to engineering in the name of cost optimization and there isn't much 'art' left in it. But look around - that's true across our society.

[0] https://en.wikipedia.org/wiki/Massing

> You seem to be making up arguments against some caricature that you have built in your mind in order to satisfy some form of virtue signaling or sanctimonious belief. Whatever it is, it makes no sense and I feel dumber for having read it.

> Go touch grass, it's healthy. (And get off bluesky)

You seem to have been triggered by the OP pointing out that a lot of this traditionalist architecture movement is - yes - racist old white dudes who want everything to be like it was when they were kids. Nostalgia is a powerful force, but it's worth remembering that the times filled with traditionalist architecture also had entire lower classes that slept in Flop Houses [0] and Shantytowns [1].

Ultimately traditionalist architecture may look nicer, but it's not surprising that buildings which cost more to build also look nicer.

If someone simply pointing the driving force behind this upsets you so much that you start posting replies like this, I would advise spending less time on X. Go outside, take a break from your "culture war". It's healthy.

[0] https://en.wikipedia.org/wiki/Flophouse

[1] https://en.wikipedia.org/wiki/Shanty_town

If you read the chain of comments again you will notice that eye tracking is only one of several things that it is doing.

> The missing part is that traditional building used to be cheap (in fact the cheapest way of building things).

This is only partially true.

Traditional building used to be cheaper than it is now, yes, but things like ornate decorations and facades were still expensive additional work. It's worth considering the survivorship bias here - the traditional buildings people know and love are the ones that were well built and cared for.

They've doing inside-out SLAM to map the environment and tracking your eyes for foveated rendering at the same time. That's fairly heavy - I think the compute required would necessitate a much higher speed connection with a desktop and a fair bit of CPU time, so it's probably cheaper and easier to put the compute right in the headset instead of scaling up the connection.

Also, they probably landed on this rough design a few years ago - before the AI hyperscalers bought up all the RAM - so the economics changed a lot under their feet.