I got to use it for a couple of days on Pro. Feature implementation actually fit into subscription usage, but then it went on chasing its tail fixing bugs it introduced and burning through about $15 in credits. It got the job done in the end, though. I'd say it's good, better than Opus, but I found that both GPT 5.4 and 5.5 are way better than Opus.
I heard it's extremely persistent and thorough. That's probably overkill for most tasks, e.g. I've heard it taking a long time and burning lots of money for tasks that could have been done much faster with cheaper models.
But for when you need it, I would love to have it.
I agree, I worked with it for a bit and it seemed better than Opus 4.8.
One thing that I would like to try is Pi code agent. People seems to like it. GPT allows you to use it on subscription, but Anthropic cut that off on Max subscription. One more incentive to try something new. Also, I feel that I rely on Anthropic too much and none of these guys could be trusted. Next step -> foray into the Chinese models (admittedly not as good). Hopefully, they will get good enough soon.
I've essentially had two subscriptions since last May (along with my use of local models for whisper, image processing, and toy experiments). And I'm about to do the same thing as you with my second account - chinese models.
I think that's important actually to have constant access to two so you can understand what their strengths are. Plus it's easier mentally to switch your alt account than your workhorse.
I love buying and reading physical books. However, about half of the books (I read mostly programming books) have letters that are printed pixelated. This is infuriating to me. No one bothers to run a trial print and see what comes out?
The root cause of this: PDF will look fine, but the text color is usually set slightly off black (why!!??). The eye couldn’t really see the difference and PDF renders smoothly. However, commercial printers couldn’t handle that properly.
Solution: set the text color to full black, you are using (most of the time) black and white printer!
You might need to have two PDF versions: one for printing and one for digital distribution (but why would you have off-black text anyway?).
> but the text color is usually set slightly off black (why!!??)
This can be cause by colour management. If the black is defined in terms of RGB and then converted to CMYK as part of the pre-press workflow, you'll typically have a mix of all four inks, and not necessarily 100% K - it depends on the colour profiles. For a black-only print job the C, M and Y channels will then be discarded, leaving a maybe-not-pure black.
>> (but why would you have off-black text anyway?)
I had read a long time ago that when doing web design you should avoid pure white and pure black, especially when one is on top of the other. I presume it’s to avoid harshness or to keep the “white” from halo’ing into the black (made even worse on a CRT display).
That is probably the worst advice when doing a printed medium, though. Different medium targets sometimes have conflicting “best practices.”
Because pure black causes eye strain. Dark gray on white is superior for long reading sessions when your paper is white. The contrast really hurts after a while if you do pure black on pure white. This is a known phenomenon.
In fact, there's experimental evidence (https://www.nature.com/articles/s41598-018-28904-x) that this high contrast plays a hand in the onset of myopia, which in extreme forms is correlated with glaucoma and other vision disorders.
> Dark gray on white is superior for long reading sessions when your paper is white.
Color is the ink's job. Approximating a lighter shade of black than the ink produces by speckling the output with tiny white pixels is definitely not an improvement in readability.
We were told by some very brilliant people (Sam Altman, Dario Amodei and to a lesser extent Demis Hassabis), that AI will shortly automate much of Research and we won't need that many researchers anymore.
Therefore, these personnel restrictions shouldn't matter, right?
I am loving the articles alternating between "software engineering is dead" and "I am going back to coding by hand". I guess we have a difference of opinions here. :-)
Judging by the amount of bugs in CC, this model can't be all that good.
But regardless, what is the point of paying to Anthropic if their models are not available to you? I am switching to GPT 5.5.
reply