noname120

> Remember that there is still quite a bit of friction to doing that, and that many people have better things to do than jump through those hoops

The thing is that Photopea has been extremely obnoxious with giant colorful ads. They take a big portion of the right-side of the screen (anywhere between 1/4th and 1/6th depending on your screen resolution), stealing important real-estate for actual work.

The owner of Photopea also fights ad blockers harshly and he went as far as surreptitiously breaking some features on purpose when it detected that an adblocker was installed (of course with a fair share of false positives), and then showed later a popup telling the user that they lost their work because of the adblocker. I think he eventually backpedaled on this due to the massive backlash.

You can pay to remove the ads but it costs 60€/year. I think few people are willing to pay for that when they just use it 2-3 times a month.

Overall the developer made the friction so extreme that people will do anything to get rid of the DRM entirely and use self-hosted instances. I think Photopea’s days are counted, a community effort to rebuild a similar tool from the ground up will definitely become a thing. The only thing left to say will be farewell Photopea, you served us well.

What makes you think it’s “BS of the highest order”?

Maybe you should read some of their stuff before forming such a strong opinion about them. And LeCun should too, he repeatedly always refused to read any of their research work and instead just insults them over and over. This is unscientific at its peek and he should be deeply ashamed of his behavior, especially as someone with such a far-reaching voice as he has.

That’s very possible and in fact until recently you couldn’t enable or disable that feature in an existing project, it was only configurable when creating a project. And for some time you couldn’t even move in and out conversations from a project configured with project-only memories/files

Just because the underlying kernel (Zircon) is designed to be good for both phones and computers doesn’t mean that Fuchsia (which is just an OS that uses that kernel) is.

Not true. When you create a new ChatGPT Chat project you decide whether it shares its memory/files with the rest of the workspace or if it should be fully isolated.

> It's also is a godsend if you use an MCP like oracle to automate the process of calling the Pro model on particularly tough problems

As far as I know Codex (at least the GUI) can automatically call the ChatGPT Chat models (including Astra 6 Pro), you just need to @ a ChatGPT Chat conversation from within Codex and tell it when to use it.

> There used to be a toml file workaround for Codex to extend the GPT context window to 1m, but this stopped working on the plans and only on per-token billing

Not true, it works again[1]. I confirm that it works both on 5.6 Sol and Astra 6, possibly other models too.

[1] https://x.com/thsottiaux/status/2089082893804896524

Cloudflare Quick Tunnels have existed for more than 5 years (yes, including the anonymous quick tunnels). I copy-pasted the url of the HN link in archive.org, see for yourself: https://web.archive.org/web/20211202005430/http://try.cloudf...

Is a new vibe-coded landing page for a 5-year old product really worthy of being on the front page nowadays? There should at least be a [2021] in the title.

How does “open sourcing the tech” “give nefarious actors on the internet more opportunities”? Sounds like a paltry excuse for not open sourcing your tech.

Misinformed comment

It prevents unauthorized access from OpenAI’s model training.

Same here on Pro 20x, Switzerland

Which OS are you using? On macOS I don’t have any of the issues you reported. It’s snappy and fast, no freezes whatsoever and sane shortcuts.

Could you elaborate?

I’m disappointed they did. M365 represents a tiny proportion of the budget especially in light of the high Swiss salaries in government (~ $100-250k per employee per year) and all the surrounding infrastructure costs. Microsoft software became really good over time and the loss of productivity and the inefficiencies introduced by the change to clunky software will absolutely not be worth it.

The elephant in the room is that ChatGPT Work/Codex use agentic quota, whereas Chat doesn’t (and is effectively unlimited on Plus and Pro plans).

On this principle I’ve built Codexify[1], a connector that provides a Codex environment to ChatGPT Chat. This enables unlimited 5.6 Sol high/xhigh usage on the Plus and Pro plans as well as access to the 5.6 Sol Pro model (which is not available in official Codex).

[1] https://github.com/devnoname120/codexify

Maybe he meant EUrouter? But it seems to be an entirely different thing: unified API for models vs a token marketplace

Hi! Any chance you may have tangential positions opening in Zürich?

Apart from the local scaffolding you can configure very few parameters, and the ones you can configure are the ones that don’t matter

Gosh I thought I had edited it to macOS 15, now it’s too late

macOS 17 is decisively faster on my MacBook Air M1 16 GB than macOS 26/27 is on my MacBook Pro M2 Max 32 GB. And I don’t mean just the animations themselves but lag/sluggishness/responsiveness in general.

Long-acting antiretroviral injection ≠ vaccine

Not true. This is the case for the API, not for subscriptions: they are hard capped at 400k.

Not true. Try to increase the context window and you will be greeted with an error when you reach the real limit enforced server-side.

Source?

Your comment is not substantiated at all.

ChatGPT 5.6 Sol Pro believes that the proof is sound. Usually it’s very good at determining if proofs are correct and their mistakes (a friend of mine is a top mathematician researcher and confirmed): https://chatgpt.com/share/6a515ead-b464-83ed-b85c-c8674f56ea...

Personally this gives me additional confidence that this is the real deal.

Given that both Gemini 3.5 Flash (high) and Gemini 3 Flash Preview (medium) beat GPT-5.6 Sol (high) for correctness and score in your benchmarks I don’t trust them at all. The rest of the ranking also doesn’t make sense, like GPT-5.3-Codex (medium) performs better than Claude Opus 4.8 (medium) yeah sure

AI fluff

I agree, it’s so much better than Docker Desktop, Podman, and Colima. And not just by a small margin, it feels several orders of magnitude faster and more lightweight thanks to its WSL2-like architecture.