Espressosaurus

Anthropic does not control the last 15 years of my digital life.

Or the method by which I reset accounts for other services.

Or my phone.

It feels like I’ve got TikTok brain for coding and have trouble staying focused on the boring grunt work of coding (like threading an API change across the required stack).

I can still do it but it requires more willpower than it used to in part because I’m getting used to telling some LLM to do it for me.

Maybe those no AI Friday ideas have some merit.

Yeah, if anything the problem is that the output uses too many words for too little signal, and incorrectly uses confidence based on insufficient information to the degree it’s clearly bullshitting.

If rent no longer rises then the profit doesn't rise commensurately with demand, so the money isn't there to build supply.

Rent control crushes supply to the degree that the rent control caps rents below the actual market rate.

If you're lucky enough to get rent controlled housing then you're in good shape. But good luck getting it when things get far enough down the timeline.

Rent fixing is the other side where supply is manipulated to extract the maximum profit. In this case the coordinated rent pricing action behaves like an oligopoly.

An experiment doesn’t say anything about user facing or not. I do tons of them when reverse engineering poorly documented hardware and APIs to determine if the thing I’m trying to do is even supported.

You need to adjust your AI detector. I read the whole thing and it reads exactly like all the Claude reports I spend all day reading when I’m using it to write code or investigate something.

It's not inconsistent if you think of it as the finances falling apart regardless of how useful the tool is.

The claims and data are interesting (perhaps overstated in the usual Claude way) but it's about 20x more prose than is actually justified.

Interesting piece, I just wish the author had presented the data and their thesis instead of making Claude vomit out 20 pages of trash around it.

Why can't they call it a prototype or experiment? Sheesh.

I'm guessing conservatively at least 10x that were sickened given how few people actually go to the doctor when they're sick.

17K confirmed cases means some multiplier unconfirmed. That is not "tiny", especially given the geographic area it covers. It's a massive failure of our food safety and yet another sign the institutions that have been systematically hobbled that keep the food we eat safe and the air and water clean are finally breaking under the deliberate neglect and abuse that's been heaped on them.

It's a lot like AWS. Yes, the uptime is likely to be better than your on-prem solution. However, unlike when yours inevitably falls over and has zero impact on anybody else, when AWS falls over it wipes out entire swathes of the internet.

Except now it's our food supply that's wiped out by the consolidation. I haven't bought bagged lettuce in months in response to this outbreak.

Smaller organizations means redundancies and options. Redundancy is the same thing as saying inefficiency. Redundancy is good.

LLMs remade Clippy. This sounds like using Microsoft Bob to manage them.

Allowing allocations to fail is a solution.

Alas, not one that was made back in 1970, so here we are.

Yeah, the lack of hardware interlocks are part of what makes it a (literal!) textbook case for software engineering ethics.

I remember talking about it several times over my college career, and for good reason.

To this day when people start talking about software in the loop in a safety context I get extremely twitchy. I don't let that shit happen on my watch.

Windows uses eager memory allocation, and applications will die when they allocate (and think image editor loads a beefy image) or fail to start when it runs out of memory.

Linux uses lazy allocation and overcommits, so what dies isn’t necessarily what you would expect thanks to the OOM killer.

With a fixed swap size windows doesn’t necessarily slow down as it runs out of memory.

YMMV if you don’t have 64 gigs of RAM and a 64 gig swap.

The latter.

Yeah, I remember an mp3 rating >50% CPU time on one of those early pentiums we had

I’m working at a company where everyone is using LLMs for everything and I’m not aware of anybody using anything truly genetic. It’s all human in the loop babysitting the tool.

I think the parent's point (and perhaps the article author's point) is that he didn't spend his entire time there. He spent long enough, but not too long before he connected with the existing body of work, but now with a way to bring his own creativity into it.

Every technology, every action has a moral aspect to it.

Just because you want to ignore it doesn't mean it's not there.

Society then decides on what morals we want to uphold, or at least tolerate, vs. those we want to change.

Laws are hard power to change behavior. Influence/goodwill/etc. are soft power ways to change behavior and bend organizations in ways they may not otherwise want to.

Would you pay a pto keep grandma from eating dogfood if it wasn't the law that you have to put a portion of your earnings into Medicare and Social Security?

The problem is one of expertise, sometimes general, sometimes specific.

If you don't know better, you don't know better to question what the AI says.

I've seen this in the work environment with a coworker who insisted that I implement my side of the control system using the control law ChatGPT recommended instead of building off the empirically tuned control law. I eventually sectioned off a part of the codebase for him to work on independently.

Needless to say he didn't get a whole lot farther.

Later characterization of the entire system end-to-end showed the existing system was already close to the theoretical limits and ChatGPT's tearup would have bought us precisely nothing except for more work to tune the new control loop.

And I see this in everything that requires expertise. You need to know enough to know when it's bullshitting you, and it's hard to be enough of an expert in everything to tell when it's bullshitting you for something you aren't enough of an expert in.

I don’t know how you stop the slop writing. What the LLM writes has the voice of whatever has been RLHF’d into the weights. There will be a voice. It will have ticks.

They can change it but it will be there.

I do like that I have a term for that “this, not that” phrasing that’s nails on the chalkboard for me after several years of reading slop.

Look at the glacial retreat around the mouth of the fjord (which is why tourists go to the town) and yeah, the amount of change is insane.

Any particular iceberg isn't that interesting, but the source! That’s the sad part.

Auto sales, airline ticket prices (notice the expansion of 1st, business, and comfort class seats), etc.

In the long term the sun will go red giant and all life on earth will die.

In the short term I still need to eat, and tech integration that takes 20 years in the economy might leave me totally fucked!

I’ve observed the degradation, but I suspect what’s happening is they’re tuning it for lower inference costs. Maybe turning down the amount of thinking, maybe quantizing, maybe something else.

It seems like there’s a week by week and sometimes day by day change in performance when on a subscription plan using their harnesses.

Machines are not humans.

Except I care about the reasoning trace because I can stop a rathole that burns $50 of tokens as it chews for 10 minutes before I get the desired feedback that it’s done the wrong thing.