monneyboi

I'm playing around with some ideas on top of pi durable. Cascading contexts exposed over http. Mostly deepseek, glm and kimi in there. https://github.com/resolveworks/key

This is taxing communication and should not be legal. I don't see how it's different from having to pay for talking to your neighbor.

I think we can stop pretending it's "pretty close" and just call it what it is.

If you want to support Quad9 in their dealing with this, you can donate to the cause here: https://quad9.net/donate/

Quad9 is a Swiss-based not-for-profit organization whose operational budget comes entirely from sponsorships and donations.

Keep up the good fight! I'm happy that you take a stand for your beliefs and try to do what's best for all of us.

I just learned the embarassing way that my ability to pay for groceries depends on the livelyness of europe-west2.

Great advertisement for deepseek and a clear signal open models are winning.

If only it would be possible to "close your doors to AI agents". It's not I'm afraid. Where do you draw the line? What browser is allowed?

For example, I now have to `pacman -S chromium` to buy tickets from Deutche Bahn. They accuse me of being a bot when I use firefox on linux. It's a known problem, and I think stuff like this will only get worse the longer this cat and mouse goes on.

We are entering? I feel like we've been there for years.

I never understood memory solutions for coding agents.

You have the whole session history right there. One recall skill and some JSON parsing gets you grep over perfect memory. Why would you ever use more tools to spend more tokens to construct a imperfect memory next to your session history?

I just don't get it.

So we pay Apple for friction. And the state pays for Graykey to remove it. Whoever wins that arm race this quarter determines what our rights are worth in practice.

Haven't touched Anthropic models ever since GLM 5.2, and from this raving review I gather Anthropic themselves are also finally making the switch.

I wonder what the future is of SaaS in a world where everybody can vibe up what they need on demand.

In China the government controls the discourse. Here it's the corporations.

Moonshine.ai

I've put it in a simple android app https://github.com/resolveworks/scribe if you're interested

Refreshing, now let's make this a default feature. I imagine a "Upcoming models" list with links to these kind of dashboards.

If we get to try and explain these times in like 30 years to our (grand-)children, I'm sure it will be with a great sense of embarrasment.

I hope one of these days we can get our priorities straight.

With every message board we find, I can't shake the feeling that this is just the tip of the iceberg.

It's only possible to get away with this because we have anthropomorphised the models to a certain extent. We can pretend they hold the responsibility. instead of the people executing them.

I've heard good things about https://bunny.net/

I use treesitter to build outlines of files and directories: https://github.com/resolveworks/trace

Have not benchmarked it, the intent is mostly to save time, not necessarily tokens. I noticed that the models need a lot of toolcalls to ground themselves, and often have trouble with getting an overview.

I can't be the only one that is getting tired of this.

Framing learning from observation as somehow bad. Something literally everybody is doing, model and human alike. It's the process this whole industry is built on. Pretending that this is bad because the other people are also doing what you have been doing, is hypocritical and childish.

Meanwhile I'm sitting here looking at some "Flibbertigittering" spinner like some sort of caveman, unable to steer the model when it misinterprets my ambigious prompt because I'm not allowed to see 80% of the output tokens I'm paying for.

Instead of this push for more vendor lock-in, give us the reasoning tokens we pay for. Thanks.

Imagine the sheer amount of power you could save by releasing the paper.

While initially a perpetual right to copying for some publishers, copyright was made into what it is today to "promote creation". Funnily enough it's used in the exactly opposite way. Culture is basically frozen, same movie and same music released every week.

The fields where creativity flows the most seem to me the ones furthest away from copyright. FOSS and memes come to mind.

What matters is who made something, regardless of who copied it. I think that in a ocean of AI generated content, this will become more and more important. Platforms like patreon and kofi show that people care about creators and are willing to support without 'buying a copy'

Why are we framing this as "the fight against climate change"? This framing puts the culprit outside of ourselves, as if it's some external force.

I feel that way of looking at it makes changing impossible. It's our own behaviours and systems that will have to change. There's nothing to fight here, the answer instead lies in collaboration.

Check the developments in ASIC. Taalas and the likes.

Anything for some more attention

It sucks that we got to "European" being used as a main selling point.

Imagine the enormous complexity of office file formats if your business is literally building coding models/agents and still you choose to bundle LibreOffice.

I'm hoping that Mozilla Firefox can adopt a similar foundation/donation strategy as Signal has proven to work quite well.

We really need independence from Google in the browser landscape if we want the open web to survive.