xvector

Am I the only one that finds the "laissez faire" attitude of the tech crowd insufferable?

There is this chronic refusal to believe that new technologies can cause harm.

My team works on testing risks from both companies' leading models and I feel pretty confident saying that uncensored next gen models becoming "open" would be a net negative for society.

I don't understand why no narrative regarding the risks of technology can land with the tech crowd until it's too late (hello, social media!)

so it'd be preferable if they didn't include the model at all?

Yup - who cares about x-risk or red lines for domestic mass surveillance anyways? I draw my red lines at prioritizing profitable customers when heavily resource constrained. That's the true definition of evilness!

If it's that big of a problem to you, you're free to just... not use the freebie?

HN needs to take a chill pill. Could it be that Mythos is expensive and they just want to give people a taste of it? I mean the alternative is not offering it at all?

They are almost certainly an OAI employee, reading their history.

For some reason I don't see you calling OAI petty when they donated $20M to Trump & worked a secret deal with Hegseth to usurp Anthropic and erase the red lines they had in place.

Starting a race to the bottom where every AI company agrees to "all lawful use" such as mass domestic surveillance and fully autonomous weapons, probably increasing p(doom) by some amount.

All to stick it to Anthropic. That's not petty to you?

To me it is an order of magnitude bigger than all of the stuff you've described. I suspect some people here just work for OAI.

Pretty sure that was a bug, I had the same issue but updating fixed it

Rose tinted glasses

...are you talking about the app? Come on. The app is for quick queries. You should be using Claude Code or Cowork.

Yes most companies do in fact operate like this. There are tens of thousands of companies that will pay more for the best thing and call it at that, because the cost is dwarfed by what even marginal gains in quality unlock for the business.

Spent a lot of time with "open models." None of them come close. They are benchmaxxed. But you won't hear many of the open model fans on HN admit this.

The open model mentality is also just so bizarre to me. You're going to use an inferior model to save, what, a couple hundred bucks a month? Is your time really worth that little?

No one working on a serious project at a serious company is downgrading their agent's intelligence for a marginal cost saving. Downgrading your model is like downgrading the toilet paper on your yacht.

The cost is so small relative to the increase. The cost whining on HN is bizarre to me. Feels like everyone here is on an individual plan and has no understanding of what margins look like for actual business.

Meta pays $750k+ TC and makes far more profit/eng, do you think they care about $5k/eng/mo in inference? A 1.1x increase would be so significant that it would justify the cost easily, especially when you can just compress comps to make up for it

I too am finding 4.7 a significant upgrade, it's hard to go back to 4.6 for me. I don't understand everyone calling it a disappointment but clowning on Anthropic is the trendy move these days.

And what's missing in all these token count complaints is that 4.7 is actually cheaper overall anyways because it produces fewer output tokens.

HN is getting ridiculous. You cannot seriously be complaining about Opus token usage on the Pro plan.

Adaptive thinking is optional

This is just a case of mass social psychosis. Claude is the same as it has ever been. Just look at historical benchmarks: https://marginlab.ai/trackers/claude-code-historical-perform...

Look where France is now. Can't afford their own retirement.

This mindset trivializes the immense achievements of "the common man" over the course of millennia.

[flagged]

Will you eat your words when major vuln disclosures come out 3-4 months from now?

And there's a reason they've achieved precisely zero penetration amongst normies.

A chat app is useless if your friends and family won't use it.

It's an LLM.

Yet more AI slop on HN

The post misunderstands how the permission system works.

Giving access to a file via the Open and Save panel is an explicit declaration of consent.

Because the panel is provided by OS itself, the app doesn't get access to the item until the user has selected a folder or file through that panel.

Just because agents aren't immune to prompt injection doesn't make it so that they aren't fantastically capable

Most big tech companies have access to the model, you can absolutely "validate their claims" or talk to someone that can.

It's really not some conspiracy. I imagine we will see vuln reports soon.

You don't need to believe it. The real story will be if companies allowed to use it, stick with it.

This is because no one bothers to set thinking to high, as it now defaults to medium in CC.

Once you set thinking to high it works just as well as 5.4 even for pretty complex tasks