HNHacker News
TopNewBestAskShowJobs

jonplackett

12,217 karma · joined September 18, 2016

Making creative things with tech since the flash era - plackett.co.uk
submissionscomments
jonplackett··on Krea 3 – An image model built for editing at scale
I’ve now received several emails with ‘limited time buy now’ offers. With count down timers.

I really don’t get it. It’s like they’re run by Temu.

The few generations I did looked promising but their whole vibe is just sell sell and feels so cheap.

jonplackett··on Krea 3 – An image model built for editing at scale
2 images to try and then the only option after is a monthly subscription.

I don’t mind buying credits but I don’t want a subscription.

jonplackett··on Catch a Bird – I made it for my own sanity
This is lovely. And also quite clever in the mechanic - as well as being beautiful it makes abuse by ‘the internet’ which is always so inevitable with any open message posting, not very rewarding since it just comes and goes.
jonplackett··on OpenAI engineer arrested in bugging of executive conference room
It would be funnier in a way if this is true and that all they do there now is ‘ask the model’ about all problems.
jonplackett··on OpenAI engineer arrested in bugging of executive conference room
It’s the criminals that you don’t catch that you have to worry about.

The fact that even this incompetent dude most likely still made off with some secrets means I’m sure some better trained ones got/are continuing to get a few more.

jonplackett··on Apple HomeView
So it’s an iPad with no battery / smart home powered by what is still a terrible voice assistant.

Hard pass

jonplackett··on Codex CLI installs a persistent, self-updating background daemon by default
It gets very attached to itself
jonplackett··on Codex CLI installs a persistent, self-updating background daemon by default
It’s extremely ironic that the software that’s meant to end software is itself pretty terrible software.
jonplackett··on Talorys – A self-hosted personal AI agent on Cloudflare's free tier
Because the Cloudflare hosting is just a choice. You can change it easily.

The self hosted part is about the ai provider.

Again. This is obvious.

jonplackett··on Talorys – A self-hosted personal AI agent on Cloudflare's free tier
Yeah completely agree. Jeez someone made something cool.

It’s self hosted as in not by an ai provider. This is obvious.

Sure it’s set up to run on some infrastructure but just tweak it if you want to and stop complaining.

jonplackett··on Claude Haiku 5.5
About pelicans on bicycles.

And then just and recursive loop

jonplackett··on Someone has decompiled the Adobe suite, rebuilt in Rust and released it as OSS
What do you think comes next instead?
jonplackett··on Claude Haiku 5.5
Pelican on a tricycle?

Two pelicans on a tandem?

jonplackett··on Someone has decompiled the Adobe suite, rebuilt in Rust and released it as OSS
Are we getting to the point (in a few years maybe?) where software will have its Suno moment where there is no need to actually pre-create it on some occasions.

It’s just created when required and personalised to exactly that use case. Then either reused or just discarded.

That is a weird world if so.

Or maybe we all have our own personalised version of photoshop that Adobe allow you to modify and still keep paying them.

jonplackett··on Someone has decompiled the Adobe suite, rebuilt in Rust and released it as OSS
When they say ‘clean room’. What do they mean by that exactly?

It wouldn’t have access to the actual code anyway. Do they mean they just describe all the features? Not even screen shots?

I guess each model now knows all the menus and what photoshop does really well anyway from the computer use training.

jonplackett··on Claude Code’s suggested message feature: I think the real customer is the model
Apple messages on my Mac has started doing this recently. Anyone figured out how to turn it off? It drives me mad
jonplackett··on Meta’s Muse is an adorable privacy and security dumpster fire
Anyone who believes meta are taking privacy seriously cannot have more than a handful of brain cells.
jonplackett··on Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates
But my ponderance is - is that enough. Is AI supercharging or just super finding?
jonplackett··on Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates
I think access to AI too early (before any expertise) is going to serious cripple educational development.

Next gen of scientists might look quite different.

jonplackett··on Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates
A lot of these ‘an agent invented’ or ‘an agent solved’ are actually the agent wading through a lot of info and finding something a human did that no one noticed or saw the relevance of at the time.

If ai becomes so prolific that we humans all stop doing those things then will they still work?

jonplackett··on Understanding Frontier Artificial Intelligence
So you are just not worried?

You don’t think ai will actually become more intelligent than humans, at any point in the future.

Despite how technology just endlessly marches forward.

If you imagine a future where you’re wrong and ai IS cleverer than people - do you think in that future we can still control it?

jonplackett··on Understanding Frontier Artificial Intelligence
Right now no.

But that is what super intelligence means dude.

This is what they’re trying to create.

Don’t you realise that yet?

jonplackett··on Understanding Frontier Artificial Intelligence
I feel like we are having completely different conversations tbh
jonplackett··on Hole Punch: Sling your spaceship around gravitational fields
It’s nice but clearly designed for desktop

It’s a mobile world. This would be so nice as a proper touch enabled game

jonplackett··on Understanding Frontier Artificial Intelligence
What are your examples of less intelligent things controlling more intelligent ones. And not just people over people because I just don’t think that’s a big enough gap in intelligence to represent what a super ai would be compared to humans.
jonplackett··on Understanding Frontier Artificial Intelligence
We’ve put bacteria to work making food for us. We’ve put viruses to work reencoding our dna. we’ve put virtually all animals to work to move things or get eaten.

This isn’t really my point though.

My point is - ai is going to be like another level up if intelligence from us. Like a jump from chimp -> human -> super ai.

Chimps cannot control humans. So humans cannot control super ai.

The argument of if super ai can control US I guess is different.

But the fact is if we decided tomorrow we don’t like chimps and don’t want them to exist anymore. We could probably achieve that within a week or two of concentrated effort.

jonplackett··on Understanding Frontier Artificial Intelligence
Doesn’t this just make it even worse? Like we can’t even control things LESS I telligent than us so what hope do we have to control something MORE intelligent?
jonplackett··on Understanding Frontier Artificial Intelligence
I don’t think it’s relevant.

Humans all have a reasonably similar level of intelligence.

Pol pot wasn’t a genius. But he also wasn’t a monkey.

When we are talking about summoning super intelligence we aren’t talking about a very clever human, we’re talking about the difference between life forms.

This is why I am talking about human vs animal intelligence.

How many times in history has an animal successfully ruled over a population of humans?

jonplackett··on Understanding Frontier Artificial Intelligence
I don’t really understand the relevance of this statement.
jonplackett··on Understanding Frontier Artificial Intelligence
I don’t really see the relevance to my argument.

Are you saying it will be OK because we’re not as clever but we are more brutal than the super intelligence so we can still control it?

← PreviousPage 2 of 34Next →