Macjam🦉

The cutest cat of all ^-^

[Macjam🦉 @Macjam0]

@TokenGremlin Merging chat, work, and codex seems like the logical step, and seems like it's already in the works. Of course, they're going to have to have one measurement for all of it.

Merging usage is not a problem, so long as you can still get the same amount done in both aspects.

[Macjam🦉 @Macjam0]

@thsottiaux I hope there's a “one more thing” on the last day moment. But banked resets are always nice.

[Macjam🦉 @Macjam0]

@WisebyIllust I'm ready ❤️

[Macjam🦉 @Macjam0]

@Niina217_Iio Thats not SFW by the way

[Macjam🦉 @Macjam0]

@HaryMik20 @SirRippovMaple @jeffwhelpley @SisyphusUnleas Most of the allegations brought up between his first and second terms were tempered or put on hold due to not wanting to influence a political candidate.

However, after his second term, he won't be able to run again, so that won't hold them back.

[Macjam🦉 @Macjam0]

@maddiedobles If it helps, I'm from the UK, and I could see both your posts perfectly fine: this and the last one.

[Macjam🦉 @Macjam0]

@kimmonismus People should start posting benchmarks where they include Claude, but the Claude result just shows a denial message.
That'll get them to change their mind 😏

[Macjam🦉 @Macjam0]

@kimmonismus Wtf… Recreating a game or product from scratch isn’t inherently illegal, provided you’re not infringing on protected IP. This feels more like Anthropic caving to pressure from big companies than any genuine legal concern.

[Macjam🦉 @Macjam0]

@Disass Wtf… Recreating a game or product from scratch isn’t inherently illegal, provided you’re not infringing on protected IP. This feels more like Anthropic caving to pressure from big companies than any genuine legal concern.

[Macjam🦉 @Macjam0]

@scottstts @SemiAnalysis_ Now, if you take into consideration that Sol is much more token-efficient than Opus, you don't actually need as much to get the same work done, showing that this chart is yet even further less representative.

[Macjam🦉 @Macjam0]

@scottstts @SemiAnalysis_ The chart inflates Claude’s “API value” by using dollar costs rather than actual token usage. Opus 5.5 costs twice as much as GPT-6.1 Sol ($4 vs. $2 per million input tokens), effectively giving Claude a 2x advantage. Normalize for token volume, and the bars get much closer.

[Macjam🦉 @Macjam0]

@kimmonismus Oh, I probably do another one of those soon, actually.

[Macjam🦉 @Macjam0]

@kimmonismus I mean... he's not wrong.

[Macjam🦉 @Macjam0]

@JPT_Arts Okay, now it's peak!

[Macjam🦉 @Macjam0]

@ChibiReviews I assume there must be some level of harmfulness to it, for them to do this in the first place.
These things generally don't come out of thin air.

[Macjam🦉 @Macjam0]

@kimmonismus @thsottiaux @StefanoGPT Because this is exactly how it should have been the whole time. The price of the plan now directly reflects the multiplier:
100 = x10
200 = x20
500 = x50

[Macjam🦉 @Macjam0]

@hotmob @kimmonismus Well, technically, one month after the backlash, because they announced harving Pro 200 last month.

[Macjam🦉 @Macjam0]

@kimmonismus We didn't only win 100 and 500 getting doubled, so it's now 10x, 20x, and 50x.

[Macjam🦉 @Macjam0]

@thsottiaux You also technically gave us all a reset, by the 50% emerging back in our accounts. 😋

[Macjam🦉 @Macjam0]

@thsottiaux I love the fact that I opened X to complain because I thought we got a reset, and I'd already used half my usage.

It was actually the remaining 50% being added from this change, because I had 0%.