jaggederest

My read, but the effective policies are infeasible, and the feasible policies are largely business as usual. Pretty much precisely as expected, honestly. I struggle to see how anything like this is going to be implemented without business as usual changing in ways that the current system is extremely resistant to.

Or encounter their CLI and it's lovecraftian combination of positional arguments, named arguments, piping, json documents, etc, distributed largely by random lot, as far as I can tell.

Yes, but it's exactly the kind of hybrid between a regular planner and something generative (writ broadly) that they were asking about. Practically speaking if you're hitting the GEQO you've already failed as a query writer unless it's a purely OLAP on a dedicated beefy machine.

I mean I would be limiting it to a very specific use case, and the flat fee would be "plus a percentage of the underlying", something like $50/mo + 20% of the AWS spend underlying.

And for that, with a simplified use case, well, they can scale up to $20k+ a month if they like, that would be ideal, and if they need a more complicated setup or enterpriseyness, migrate off with my blessing and available-not-required hands on support (again for a reasonable fee, maybe $10k if you want the white glove).

Well, one of the things I would be very, very focused on is LLM optimization, so that is something that could be mitigated, at least. Especially since I would be building with it, certainly, I would try to engineer it so that it was to Mr. Claude / Mr. Gippity's taste

No, it's because gasoline is not a very thoroughly refined chemical, so it basically has a bunch of aromatic rings, benzene, and all sorts of monomers floating around, that rapidly (~year or two) turn into varnish and other polymerized nastiness when they oxidize and crosslink. I believe gasoline was originally a refinery byproduct that was used because it was so cheap as a fuel.

Diesel and other heavier fuels generally are much closer to aliphatic hydrocarbons (straight chains, not aromatic) and so are much more storable, because aliphatic hydrocarbons have much less opportunity for oxidation (they're like saturated fat, i.e. much more stable).

You can get polymerization inhibitors to put into gasoline if you want it for long term storage, with the caveat that they produce some nastiness at the bottom of tanks that you'll have to avoid when pumping it out, and I think they can also slightly reduce the octane rating.

If they're region-locking data appropriately (i.e. for people to comply with domestic storage / gdpr-style requirements) they really can't. Bahrain is only about 300 square miles / 80k hectares.

They could, of course, open other regional data centers in other countries, or say "data in this geo zone may be in any of X, Y or Z" countries, but for the latter that pretty starkly limits some of the major customers they'd have, I would guess, and for the former, well, they have other geo zones already, so if people weren't replicating to them, I'm not sure why adding me-east-1 me-west-1 me-central-1 would fix that issue, they just wouldn't replicate there either.

If I were a bit more bloody-minded I would launch a service for vibe-coded apps that, under the hood, did everything "the right way" and just charged a flat fee + percent on the underlying.

I feel like if this was done correctly it would eat a bunch of the market, but I question how many people are actually willing to pay for "the right way". The last time I had that experience it was with Heroku which was quite a leaky abstraction.

This is about to bake your noodle:

https://www.postgresql.org/docs/current/geqo-pg-intro.html

Where can I find an oven made with entirely US parts (or even ~85-90%+)?

I'll tell you in advance that I haven't been able to locate one in a few years of looking. Most major appliances are, at best, 60-80% american with no accurate accounting, even if they're "assembled in the USA". Cars are the only thing with accurate sourcing labels due to legislation.

> As I work in english and use them in english, I wonder if using LLMs in a different language to code renders a different result as well. Like, if some of these benchmarks were made in other languages, would the result be similar.

I am not a Chinese speaker but my understanding is that all of the models have substantially different behavior in Chinese, to the degree that it's kind of like a second model. Would be interested in hearing more if anyone has direct experience.

Makes a bunch of sense, appreciate all the detail! Similar for myself but entirely visual, so a fascinating amount of overlap.

I guess I'm wondering about the fidelity, but I'm not sure what exactly I meant. Copy-of-a-copy kind of effect, basically.

I feel you, in the same boat, though I can't actually isolate any trauma or anything like that, reliving unpleasant experiences verbatim is no fun. And the early 2000s was not a fun time with all the pop earworms running around.

It's funny, I've never met anyone but my family who can do this like I can, so cheers! I always say I don't have a perfect memory, but I don't know anyone I can test myself against except my mother, who can "relive a trail ride she took 45 years ago down to the plants and opening and closing the gates". Having to remind people about significant shared life events is a real downer.

I spent an enormous percentage of my life very frustrated with other people because they couldn't just see the answer to things.

For example, I can put an engine into a car, and still tell you where every part of that engine is located relative to every other part, without being able to see it, including the internals, and describe, for example, over the phone how to troubleshoot it without being able to see it at all in person - just a mental model. And, for reference, I have almost no fundamental knowledge or experience with engines besides taking one or two apart - and it would not extend to other engines. It's been at least 20 years and I could still do it verbatim.

I'm way out on the end of that bell curve, but I'm not unique. I've worked with people who found that ability completely baffling, and would have to break out diagrams and blueprints to do that.

The gap is a lot wider than you think it is, it's not subtle, and this is known as the typical mind fallacy.

Edit: And funnily enough, further down there's a comment about Feynman realizing that his friend counts visually and he counts verbally - and I'm on the verbal side on that one, unless it gets complicated, and then I end up tallying mentally - but I have to count them verbally at the end.

Here's an easy test: If you imagine something, is it more or less vivid than your dreams?

Either way, you've established a difference in subjective phenomenology and there are people on either side of that divide.

Drawing? doing math around how many edges are common between the two? I turn it and count, but I could also do it other ways, including exploding it into flat shapes or conceptually understanding the nodes and vertices of the graph, or counting the edges and dividing by two since there are 6 total faces.

Can you draw them and then look at the drawing? I'm just wondering about the "bootstrap" ability, for lack of a better term.

Wife is the opposite - I still surprise her when I shave my beard without telling her. She doesn't seem to have any trouble visualizing though.

For reference, she didn't know the twins from "Property Brothers" were twins until I told her, after watching about, say, 6 seasons together.

Same but in reverse, a certain psychiatric medication turned on an internal monologue I've never had before and it drove me absolutely, literally insane until it wore off. Never again. The most inane, moronic commentary about my day to day life as it happened, like Barney speaking through me. I normally experience the world wordlessly, mostly visually, and construct ideas and actions that way (thinking about, say, getting a glass of water, I'll picture the process), so having internal speech suddenly thrust upon me was awful.

it was ruby, so I did in fact know ;)

Edit: the customers, in some cases, did indeed run interpreters that were bad, but that's not my lookout! Buy me a fancy beverage and I'll tell you a story sometime

It's much easier when they release huge PR whitepapers about how many felonies they've committed recently, I think. Makes the civil suit much more straightforward in my opinion.

Appreciate the references, I was the New Relic Ruby agent maintainer in 2010, as an additional reference. And we called it an agent, because it was hosted on the customer's infrastructure, but they didn't maintain or control it, we did. Regrettably no license to kill, and not much in the way of secrecy or gadgetry.

Completely agree that the world is full of absolute nonsense, but legally determining responsible parties appears to be relatively straightforward (but does involve a judge or jury)

Like a nice chocolate bar, but echoed - so I can remember what it tastes like enough to want one, but not satisfy the urge to actually have one. Ephemeral, rather than satisfying.

Agent is an old school way to say "constantly running daemon to accomplish some specific purpose", as someone who wrote a lot of performance monitoring agents. AI is using the term appropriately, but I think assuming it's always AI-first is slightly tragic.

> If people have an "inner taste" or "inner hearing", why bother eating a sweet or playing music?

Because the interior version is strictly inferior to the real thing, at least for me. It's, maybe, 80% as good, but that last 20% is a real humdinger.

Especially for music, I have a hard time hearing it in my head and not playing/singing/drumming along with it. Currently working on "Winds of Change" by the Scorpions as practice, and it's basically been playing in my head for the last 72 hours.

I do have to listen to it occasionally to refresh, but it certainly makes learning music easier... if not infuriating. Looking at you Lou Bega, and Rob Thomas.

High five brother, I find these threads fascinating, because even within the group of people who have hyperphantasia like this, our experiences differ wildly.

"Describe smells by how they look" - that's synesthesia maybe! A bridge too far for me - I'm more likely to describe what the situation I was in when I smelled a given scent most memorably, rather than how the scent looks.

I also explain things verbally by referencing an internal picture and describing it, and find tacit knowledge very difficult to express in that vein.

Me here, inner dialogue frequently, inner monologue never (self-directed speech like "okay now I need to pick up the ...")

My inner dialogues almost always come with full fidelity, visual, audio, tactile, I can still smell the Marlboros on my dad's flannel even though he stopped smoking 20+ years ago.

The person who sets up that robot is the person responsible and effectively signing the contract by those actions.

> Who are you gonna sue then?

The person who set up the robot.