Yep, and the price point theyre looking at is 95% of an engineer.
Once they get people hooked, deskilled, and paying, the money ratchet only tightens.
And the companies KNOW that theyre replacing engineers, or trying to. So each engineer replaced is X salary a year they now have available, so make it back in SaaS LLM tokens.
I use it on Pro and was just thinking today, there is no way $20 covers the cost of it. But I'm long term unemployed and can't afford any higher tier, so if they drop it guess I'll have to find a non-anthropic solution somehow.
Sonnet in the Claude Code harness is hard to replicate out of the box. That vertical integration is not easily replaceable but by no means impossible...
OpenCode and their Go plan will get you close if you're willing to put in the config work.
For when you do need the larger models Fireworks has a pretty generous 'Pass' that comes out to about $7 a week for some of the larger bleeding edge models.
Other than that Codex's $20 plan is still somewhat valuable though they keep reducing usage. Google's $20 plan will get you some Opus usage in Antigravity and a generous amount of Gemini. Not sure how long that will last as they've been tweaking pricing and planning language recently too.
If you aren't significantly overweight a doc won't want to prescribe a GLP-1. Because if you die while on it for any reason, and your family sues, the expert witness will testify that a GLP-1 wasn't standard of care for a patient with your presentation. Then the doc's insurance loses the case (and he pays way more going forward assuming he can practice at all).
This is an argument that the litigiousness of American society incentivizes suboptimal health outcomes from the legitimate health care system because the employees of that system need to legally protect themselves; therefore a person might be better off if they go outside that health care system and buy grey-market Chinese pharmacy peptides labeled not for human use.
The engine needs to save the RNG seed too and various other details, the goal is definitely to make it as deterministic as possible (and yes saving the packets is part of that).
Screw AA for doing that spotify stunt - spotify disabled all API access except for paid accounts due to that. AA's greed and mass theft ruined a good thing.
Who's greedy? The people who put something out on the internet for free, or the people who stopped letting people access that same thing, unless they paid them money?
"But Spotify pays the musicians!" some people might scream. While true, it is a pitiful, insulting amount.
The example that hit me the hardest was Dr. Dre, whose albums were streamed 1.5 million times and earned him $4345.
1.5 million is a platinum record album by industry standards in the US, UK, Canada and France.
$4345 isn't even enough to pay a month of rent in New York, where the court case was done.
Anna's Archive is not the ones you can blame for greed.
I wrote a Spicetify plugin to show information about a song’s genres, but unfortunately the public API endpoint it was using -- as well as many other endpoints -- was removed after their changes, even for Premium users.
They do stuff like that. They also killed "Robot" personality last year which was my favorite. The replaced it with "Efficient" or something, but it isn't the same. Robot was terminator-esq, appropriate for the new age we are entering IMO.
Ehh... can't really hit "chatbot" limits on the $20 plan. Pretty sure the limits are not token based for that in the first place, and if it spews out a ton of stuff, it takes me longer to go through it and I end up asking it follow-up questions in a way where it replies... _relatively_ concisely. Still, gimme robot back. On a good note, it almost managed to call me stupid.
Codex has also been fine, but I'm guessing they know better than to tweak it like that, given their target users.
I have hit chatbot limits with the $20 a month plan. During the day I use it with Codex and I night I use it to study Spanish. I don’t know if the two are correlated.
But then I just switch to another OpenAi and strangely enough, chat forces me into “thinking mode” when that happens and won’t let me do instant
Anytime I use a system prompt that attempts to make the LLM “robotic”, I get denied. I have tried on Grok, ChatGPT, Gemini, and Claude. When I ask the LLM why, it says it’s because such system prompts are used to bypass guardrails. Maybe that’s why? idk
Yeah, I miss "Robot". It helps to add something along the lines of
"Tell it like it is; don't sugar-coat responses. Get right to the point. Be concise." to custom instructions.
reply