The money is in selling compute not using it. Anyone who thinks otherwise is a mark.

Because they're incredibly ugly

It's gorgeous and the world doesn't have enough of it.

I was gonna say, you can dislike it because of the design, but understand that that's a subjective opinion that has nothing to do with whether AI is awesome. I actually like these designs, personally.

Until it deletes your home directory, which i'd argue is an alignment problem. Destorying my data is not in line with my priorities.

Lots of people have deleted their home directories by accident. What you consider this an alignment problem?

It's an alignment problem in the sense that it demonstrates the principle that today's AI systems cannot be trusted to reliably work towards the goals of their users. A small-scale alignment failure and a large-scale alignment failure are the same fundamental type of failure. Typically, large disasters come after smaller disasters which foreshadowed the disaster mechanism, but weren't taken seriously.

I think there is a meaningful difference in kind between an AI that makes a mistake and an AI that is actively malicious.

An AI which is mistaken about the goal you give it, or how to go about achieving it, will behave in a de facto malicious manner, as this incident illustrates.

How manypeople have deleted another user's hone directory, though? That's s the proper analogy IMO.

Of the people who primarily use other people's computers, I'd assume the percentage is about the same.

Give the AI its own computer and it will not delete your home directory, because it's not actively trying to hack you.

I personally will not use a provider who uses llm tools. I know it makes me and my coworkers less careful and lazier. Qualities I dont want in a health care providor.

Ive actually moved primary care physicians over this once already, found the oldest guy I could who barely knows how to use a laptop but spends a bunch of extra time with me.

Funny how we assess risk. Not criticising you - I'm as irrational as the next guy - but "I found the oldest doctor I could" seems like it has a different set of risks baked in.

The older the doctor the more experience they have, but also their knowledge is more outdated (how many keep up with medical journals?) and their brain is worse at learning and connecting dots so I'm not sure I'd choose the oldest I could find.

The human body is the same as it was before your doctor was born.

Medicine however is not the same. Nor is our understanding of the human body.

For example, a 60 yo doctor would have been born before the first heart transplant, the recipient lasted 18 days. Now 5000 are performed each year and after 5 years 80% of recipients are alive.

The Earth was mostly the same in the 1300's. Let's use material from that time to do engineering.

Our knowledge of it very much is not

My current provider shows me their computer monitor for the entire appointment. I didn’t ask for that treatment but I really appreciate it.

When LLMs stop helping me with my work, I’ll stop going to doctors who use them.

Too bad for me, it’s a real mixed bag. I need a doctor who is an AI-using LLM skeptic, I guess.

This translates to "I'd rather get the traditional types of medical errors instead of the modern types of medical errors."

Better the devil you know ...

Likewise. I basically

- will not see a “provider”. A doctor (MD/DO) only please.

- will not see a doctor who uses LLM tools (or therapists for that matter)

- Make it abundantly clear that if you do then we’re not a match.

- Will pay significantly more/go out of my way for for this.

Regardless of whether your preferences are rational or not, you're quite lucky to be able to pick and choose. There is a shortage of clinicians and it's only going to get worse so most patients will have to take whomever they can get, especially if they're not able to wait months for an appointment.

Indeed, I’m definitely in a privileged position.

In a similar vein: Will not go to a doctor who is a DO and not an MD.

We all have our biases.

The bun route == spending 100k for an LLM rewrite as an advertisement for the LLM company you sold your soul to.

The literal worst time. I prefer meritocracies.

You think aristrocrats are using GPT and succeeding?

No I think aristocrats are using llms to justify the flattening of wages, where they can use it as an excuse diminish the value of merit based systems.. Meritocracy / education was one of the few ways someone from a lower class couod climb the ladder and they're trying to destroy that ladder.

Then that's a tirade against capitalism - don't direct it at AI. If you don't like the situation, push for socialist policies.

Use the ticket system built into mininote.ink 's mcp server. Works perfect right out of the box. Also great notetaking app.

Docs:

https://mininote.ink/docs/mcp-docs

Are you people seriously this dumb? Have you conwidered that all of these benchmarks are trained into these models. Can you stop sharing them as if they matter?

What you mean.. The tools are all just invoking bash and terminal/cli cmds and http requests. Paradigms that have existed and stayed mostly unchanged for decades.

These do make up a huge % of tool calls, but I don't think these make up a huge % of tool call failures.

I see models fail on tool calls that involve API requests to a specific API, internal or cloned Makefile calls, npm run commands, etc.

Gemini 3.5 flash is better than fable at tool calling. Tool calling is probably one of the easier things to do post training for.

I don't use Gemini cli because they're so bad at agentic work/tool calling. I use their chatbot all the time, though.