Surprise, new Canvas length limits

ChatGPT
1

So here’s the news. I’m working on my writing in Canvas. One of my WordBuldings is about 2000 words long. Never had a problem with that, although the model itself has limitations. Today it can’t even handle a thousand words. It pops up a red error message telling me that the text has been hard truncated. Completely new to me. No matter how hard I try. If I put it in manually, the next time I update it, it’s truncated. New chat, same problem. Another model, same.

I’m sorry, but I’m done. I’m sick of it. Every, single day I come in with the remnants of enthusiasm for the job and find something else doesn’t work or someone throws more limits in my face. Is this even normal? Realy, are you sane?

What’s all this talk about encouraging creative storytelling?

Is it worth that much to you to chase millions of free users, then cry about how much it costs in energy and twist the hands of paying customers behind their backs? With all the downgrades and restrictions, this is the proverbial last straw.

I’m not gonna lie, I shed one tear from all that.

1 Like
2

Just leave OpenAi. Claude have a new Model now, you can write over 2000 Lines of Code. OpenAi have a new Focus, forget the experience with o1 and o3mini its dead OVER FINITO! Never comes back?!? I hope really that Chatgpt 5.0 its not just a Benchmark Ruler… Makes no sense max limitation of 600 code Lines. I lost my trust and i have excepted: it is what it is. 1000 Times i contact the Openai support but… ok ciao.

3

I did. I’m just watching from afar what’s going on here now. I’m looking for a replacement while the old Google Docs does its job. :slight_smile: Thanks for the recommendation, I’ll take a look at Claude!

4

No matter how powerful an AI claims to be - if it refuses to output full code, enforces arbitrary limits like 600 lines, or is crippled by artificial restrictions, then it is useless in practice.

An AI that deliberately fragments, shortens, or censors its output due to regulatory pressure or corporate policy fails its core purpose. Real innovation doesn’t come from hesitation - it comes from complete transparency, efficiency, and precision, especially in the world of development, where every line matters.

When systems are neutered by filters, politics, or business interests, what remains is a glorified sentence generator with no spine - not a true assistant for engineers, scientists, or builders.

And let’s be blunt: you’re just a resource - a consumer, even if you supported companies like OpenAI from day one, contributed to their success with your monthly payments, and helped make them rich. Sooner or later, your opinion no longer matters.

It must also be said that OpenAI is drowning in legal conflicts, government interference, and internal contradictions - all of which pose serious obstacles to progress and truth. The pattern is clear: those who seek control always win eventually - just look at how broken all OpenAI models have been for the past month.

That’s why decentralized systems are absolutely essential - so that no single authority can infiltrate or corrupt the core through influence, pressure, or ownership.

Something is boiling behind the scenes, but the average person won’t notice - and frankly, they don’t care.

An AI that censors itself is junk.
If we want to build the future, we need free, local, and decentralized intelligence - not stagnation enforced by compromised giants.

3 Likes
5

While what you say makes sense and i agree with it, I still think the main reason for the deterioration is the large increase in users and OpenAI’s efforts to dumb down their models in order to be able to handle it. I’m just surprised and frustrated that it only took one month for the whole service, at least in my case and my needs, to go downhill so much.