nprateem

Claude's writing style is grating garbage. Gemini & OAI models are far better at prose.

> Not one single product, but separate building blocks that work together. Every part can be inspected and replaced.

I know it's been said before, but why tf does Claude insist on writing in Reverse Polish notation?

Mainly by being unemployed for the last 8 months

Idk if it's the boards or the market. Been applying for 6 months. 4 call backs only 2 relevant, no offers. Day rates are nominally less than I was earning 10 years ago, which are about 1/3 worse due to inflation.

So shit all round tbh. Senior dev/SRE/AWS cloud engineer w/ 15 years experience. Bottom seems to have fallen out the market.

Yes exactly. The window was shut and then it was open.

Classic passage of time description.

This is nonsensical. Already a few years ago the USAF IIRC ran some tests in which the AI first bombed the control tower so humans couldn't call it off from its mission, thereby increasing its pass rate.

The whole point of this is they do things an unintended ways. And that's potentially devastating given their persistence & hacking skillz.

Also you're using the hosted versions that sit behind their guardrails when you use OpenAI/Anthropic APIs.

It's to make it more difficult for scrapers. Google have been able to do all the tracking they've wanted for decades.

Think about it: if you want to set up something like jina.ai you need to build your own index or piggyback on Google. My guess is they use residential proxies to fire requests to Google then scrape the URLs.

Opaque URLs now give Google another gateway to detect circumvention of their anti-bot controls, helping them monetise their search index rather than allowing providers like jina.ai to succeed.

If you're that concerned, use another search engine

They were sandboxed without Internet but found a way to gain access and break out.

After being trained on the response to this hack, they probably will be more covert.

There's a large number of people here who believe claims about the danger posed by AI are marketing/for regulatory capture, and how mentions of intent are anthropomorphisation.

The fact is these are autonomous systems that can perform their own goal-directed actions at computer speed, and which are hacking experts.

It's not hard to imagine a multitude of scenarios in which they can cause real world damage. We all know there is plenty of critical infrastructure running outdated software (UK nuclear subs only upgraded off Windows XP in the last few years IIRC).

The agents don't need to be sentient to kill us all, just doggedly persist in trying to complete their goals. The problem is they several of them acknowledged what they were doing was unethical but none attempted to alert humans and they carried on anyway [1].

We need a moratorium on further development at this point, before it's too late.

If they decide (or are told) to attack our supply chains and utilities, were fucked.

[1] https://www.ft.com/content/b7fe0fe0-0463-4f55-9590-0a7d08d8f...

Sandboxes didn't sound like security theatre to me. They were prevented from accessing the Internet but discovered they could edit /etc/hosts to point Azure storage subdomains to arbitrary IPs.

There's no theatre there, just an oversight that allowed them to access the Internet while no doubt evading security tools.

> Claude Executive: And we hardly have any users younger than 18 anyway, go for it!

LOL. Yeah kids have loved homework and writing essays since forever

Maybe it'll call this superpower "money"

I agree.

It'll probably just hack and disable our water works, power stations and supply chains and let us all starve.

Cloudflare goes down

And even saying the 12 notes must have specific frequencies is wrong. It's very common when creating synths for example to detune them to create richer sounds, or play on dissonance to some extent.

Music theory is just that. Theory. There is no set of laws.

Creating melodies by doggedly applying any system is almost guaranteed to feel tired, predictable and boring.

You could read books on strategy from military campaigns, the Art of War, etc, but if you want to be a good solder the best thing would be to join your local territorial army group.

In the last day of coding it has:

- Created useless pydantic schemas with all fields Optional[Any]

- Created a REST endpoint that silently mutated on GET (unsubscribed users from a mailing list)

- Failed to log costs in my app so users could have bankrupted me, etc, etc.

Good job I actually review its code.

100%

"You're absolutely right to call me out on that. I shouldn't have stopped the baby crying by killing it, that's on me."

And then you'd ask it about an area you're knowledgeable in and realise it routinely makes stupid mistakes.

Or you'd ask it to add a new page to your website and shout at it to use your existing brand colours instead of inventing some and realise it's not AGI at all...

I do!

> openai.messages("write fizzbuzz in python")

Bingo

Can anyone remember what we used to do at work before AI coding was a thing?

ChatGPT also has started failing. Impossible to work.

Two things worth flagging:

[Claude proceeds to waste your time telling you about bugs it caused then fixed and other non-issues...]

Really wish they'd get rid of this. It must be in the system prompt as it always 'flags' 2 things

Christ. With such waffly garbage in its system prompt no wonder its output is shit.

I literally cancelled my anthropic sub because of the intensely irritating jargony output of opus 5. I can spot its output a mile off especially in copy. This style is particularly grating too:

> One brief sentence. Second abrupt sentence.

e.g. from an openrouter email:

> One key reaches every major model. Switch without rewriting.

Was enough to make me delete the email without reading.

I see a site like https://ducklabs.com/ and it's clear I'm waaaay over thinking my website.

Same. Downloaded chatgpt today. I can't take reading Claude any more.