Posts on X before Jun 28, 2025, 7:30:54 AM UTC
i've noticed a nasty effect of AI that targets people who enjoy reading
before mass AI gen'd posts/replies, it was a joy to see what people of all intellectual levels think about xyz (at least for me). every post/reply I read, I would receive dopamine
but now... 1/?
why work on airplanes, their design and usefulness was perfected 36 years ago
"what do you think they have now" things you wouldnt even know what youre looking at even if you saw it (youll never see it if its being used against you)
even drones are almost done being perfected
so for $200/month Claude Max plan, I calculated with a fine-tuned ccusage, that I used $12,000+ worth of Claude Opus 4.
if you dont think Claude Opus costs them fractions of what they claim/what they charge for it via API, idk what to tell you
they've dramatically reduced cost
3 years later after the first complaint about the same exact thing
sometimes i forget how much raw intelligence gpt 4.5 has. sure, it's still unusable for anything production oriented, but every occasional back-and-forth I have with it through ChatGPT is a stark reminder of the models to come from it this year.
want to see what it's like using Gemini CLI? Here you go:
>It seems I'm still having trouble with the file. The build is still failing with the same errors. I'll read the file again and very carefully correct the syntax. It appears I'm having an issue with mismatched braces.
i remember telling friends "you'll be able to buy bitcoin on paypal one day" and them laughing about it. a mainstream corpo, one that seemingly has everything to lose and nothing to gain against crypto, adding it to its billion-user platform?
and yet...
gpt 4.1 is near-instant recently. not talking about -mini or -nano btw.
after using Gemini CLI for 30 minutes straight, I'm now back to using Claude Code.
data -> nom nom nom -> write like @paulg please
lots of birdies talking about Gemini CLI today
Agreed. not transitioning into Dia until these 2 are done.
daily morning routine
>open @diabrowser
>still no vertical tabs
>close dia
the state of apple intelligence.
once models realize that chances are the user they are talking to is wrong about most things, thats when the intelligence liftoff happens
the state of iphone, ios, apple. amazing.
o3 1-shots often, but that's the thing - it can pretty much only 1-shot. the moment you start asking it to do any follow ups, it's over. definitely not a back and forth agent
the opposite is true with sonnet 4 and gemini pro 0605. and that's not necessarily a good thing either
the moment @OpenAI @OpenAIDevs makes the $200 pro plan include unlimited Codex CLI with o3, I'm dropping Claude Code Max
this seems like a no brainer to me? why wouldn't they do this? honestly I'm shocked it hasn't happened yet ESPECIALLY NOW with the insane price drops
so @windsurf_ai is literally unusable. for the past 3 months, every time I check back into it, I see only gpt-4.1 and sonnet 3.7. no other models - even though they advertise 10+, including o3.
i tried re-signing in, completely re-installing, asked on Discord, nothing. lmao
looks like you can already use o3 for 1x request via @cursor_ai - wow, they are on top of their game. big props.
something's definitely being updated or whatever at openai rn, all the models are dead slow, especially via API
when gemini works, it works great
but when it doesn't, it tries to gaslight you into thinking you should use svelte for your obsidian plugin. LMAO
imagine looking at this as an exec - as all the execs along the way to being green lit for the once-a-year release - and thinking "damn, this looks amazing! theyre gonna love this."
100% switching to Pixel this year, and a new non-apple laptop running arch next year. 10000%.
throughout history, competition in certain markets has progressively become harder due to global education and exposure progressively increasing, = competition++
i think we're entering a time now where while global exposure is skyrocketing, global education is plummeting.
personally i use claude code to research and then cursor to make the edits. iykyk
fyi gemini pro 2.5 06-05 works much, MUCH better when you turn off all MCPs. this is true for all models but particularly the new 2.5, idk why.
ok google I kneel
i think this trump v musk situation is just 2 guys busting each others' balls, except the way that they do it is very different from the way that you or I would do it
they're (factually) billionaire power moguls who care about their public image. so... that's what they poke at
seems like Cursor is actively working on the background stuff when it comes to the new gemini 2.5 pro model, it's definitely showing some promise. still has some issues with edits here and there but I can tell right away it's an incredibly intelligent model. low yap. smells good.
so uh why is the latest gemini model literally unusable anywhere and everyone is scrambling to write god prompts for it so that it at least somewhat works?
guys it's over pack it up go home there's nothing else left
the real eye opener here is not whether this is a real leak or not
it's that we're even considering the possibility of this being real. that's how bad apple has become
we can no longer tell apart satire from reality - and that's when it's time to realize shit is BLEAK.
i just woke up from a nightmare
i just graduated after 4 years at a top known-for-programming university and entered the workforce, and realized that all the entrance jobs are now closed because the PMs can just vibe code with Cursor instead
this used to be so fun and cool because you know that a person somewhere in this world typed out every character, thought to themselves "how would a pirate say this, how can I type this in that tone and style?"
nowadays this is just a quick prompt update and AI does the rest
new Google SOTA (and in general SOTA) dropping :)
1000000000% they nerfed Claude Code into the trash. So sad but expected, not surprised.
the problem with today's LLMs is that it generally feels like they are... lazy. to put it into one word.
surely openai's codex via the cli will allow you to connect your chatgpt pro license to it, and use codex/o3/o4-mini free of charge just like anthropic's claude code currently does with their $200 plan
if they don't that's such a user drain for them imo
i swear, something is wrong today with sonnet 4. or claude code. producing straight trash today for some reason
i wonder if there's been a silent update or something. 1.1.0 of claude code doesn't show any meaningful changes but then again idk if they share system prompt changes
i finally understand why some people follow like, 15 accounts
everyone's rage baiting, engagement farming, slop posting, for that juicy X paycheck (or hope of it)
"Claude Code."
"Claude Code."
"Do you trust your own code?"
"Claude Code."
"Have you ever committed without asking it first?"
"Claude Code."
"Do you always tell it to ultrathink?"
"Ultrathink."
"What's it like to push code that actually works?"
"Ultrathink."
this makes me feel so much better about my frequent 3 hours of sleep days
boggles my mind why people are paying for anything Claude API when Claude Max exists which gives you unlimited Opus 4
Don’t be mad at them, they had their copyright/marketing done BEFORE Codex, Jules, Claude Code, etc
Why does this simplification make me bullish for Apple
if deepseek releases something tonight im gonna lose my mind. i just started trying to go to bed at 9PM, but these last few weeks have been keeping me up with releases and trying them out... tragic for my health but (hopefully) worth it in the long term?
There was a dislike button on TikTok a few days ago, only lasted there about a day or so. I miss it already. It felt so good clicking it and actually seeing less of whatever I thumbs downed.
Idk why they removed it :(
this is done by design btw, anthropic is going all in on claude code / its flawless, never erroring/rate limiting experience
why is X starting to feel like it was vibe coded over a weekend
You can only DM on business days.