Posts on X before Jul 28, 2025, 6:33:08 PM UTC
it's so over
it's me. i'm the "one user"
nvidia will be first company to reach 10 trillion idk maybe even before 2030? x.com/AndrewCurran_/…
i remember watching nearly 2 decades ago, him being so calmly confident and matter-of-fact, all while polls against his predictions (of the top 1% researchers/industry leaders of the space) were all >80% against his predictions
he never batted an eye, never panic-explained.
looks and feels like a gpt-4 moment for video models
The introduction of Runway Aleph, our state-of-the-art in-context video model setting a new frontier for multi-task visual generation. Chat mode now available on mobile. And our weekly community spotlight featuring incredible Act-Two creations. Get caught up on what happened This… Less
The introduction of Runway Aleph, our state-of-the-art in-context video model setting a new frontier for multi-task visual generation. Chat mode now available on mobile. And our weekly community spotlight featuring incredible Act-Two creations. Get caught up on what happened This Week with Runway.
remember those videos where people would inpaint robots over people doing things? fooling boomers & co. on facebook, etc?
is there much sci fi left to fake anymore? is there anything i could watch as a faked video when i'm 60+ yo and get tricked thinking it's real?
if you think chatgpt psychosis has messed with people's minds so far
wait until these new storytelling mind-rooting models hit mainstream use
Once you get to intermediate / advanced level in Claude Code you start to see how there's very little tutorials on tehse topics and the internet is littered with beginner tutorials all teaching pretty much the same thing.
this is all you need to see/know about the new subagents mode of claude code
this ran start to finish for roughly over 1 hour
the result is perfect, chef's kiss, cleanest code i've ever seen
now imagine, i have at least 5 of these running at all times
wbu anon?
literally the only problem with claude code now
is tok/s
im forced to use sonnet if i want to get a feature done using my subagents flow
but damn is it worth it because so far it has a 100% success rate of whatever i throw at it
o3 in chatgpt feels very different
feels like intelligence is oozing from it with every word i read
what did they do?
re. claude code subagents update;
commands are basically useless now, much better results with "run my _ agent to _"
new memetic virus installed
vibe coding with claude code is basically just note-taking but your notes actually do something for you and you can run them
if you're self sufficient (self employed), not knowing this ISN'T bad
but if your livelihood depends on being an employee, not knowing this IS BAD
if i asked an llm, it not knowing this makes it a bad coder llm
tldr; treat hiring employees the same as testing out an llm
it will literally become mainstream logic of "if xyz is so smart, why aren't they going viral on tiktok about what their plan/values are?"
Zohran is the first politician to do basically this. Culture flows from the cities outward. Today you can become mayor-designate of NYC if you poast well enough on TikTok. In a few years, governor of Ohio or perhaps superintendent of Kearney, Nebraska Public Schools … Less
Zohran is the first politician to do basically this. Culture flows from the cities outward. Today you can become mayor-designate of NYC if you poast well enough on TikTok. In a few years, governor of Ohio or perhaps superintendent of Kearney, Nebraska Public Schools x.com/a16z/status/19…
crazy to see people still using llms to make single file manual edits, giving the llm no other context, just raw dogging + guessing + duplicate code must be insane
it's over for all my competitors, past future present
latest claude code update is godlike
july 2025 is about to close out,
seems like it was dominated by china
i think i'll start an end of month usa vs china thing, seeing who won the month in the world of AI (when it comes to my niche llm/coding usefulness scale only, not really keeping up with image/video gen)
🚀 We’re excited to introduce Qwen3-235B-A22B-Thinking-2507 — our most advanced reasoning model yet! Over the past 3 months, we’ve significantly scaled and enhanced the thinking capability of Qwen3, achieving: ✅ Improved performance in logical reasoning, math, science & coding ✅ … Less
🚀 We’re excited to introduce Qwen3-235B-A22B-Thinking-2507 — our most advanced reasoning model yet! Over the past 3 months, we’ve significantly scaled and enhanced the thinking capability of Qwen3, achieving: ✅ Improved performance in logical reasoning, math, science & coding ✅ Better general skills: instruction following, tool use, alignment ✅ 256K native context for deep, long-form understanding 🧠 Built exclusively for thinking mode, with no need to enable it manually. The model now natively supports extended reasoning chains for maximum depth and accuracy. Hugging Face: huggingface.co/Qwen/Qwen3-235… or huggingface.co/Qwen/Qwen3-235… ModelScope: modelscope.cn/models/Qwen/Qw… or modelscope.cn/models/Qwen/Qw… API Doc: alibabacloud.com/help/en/model-…
with standard they matryoshka doll package every item in your order
claude code has...
checkpointing now?
by the way claude code sneakily added in sub agents last night during the doc outage
weird, claude code docs are down
docs update? gui for claude code dropping?
sonnet 4 wants to show off how amazing of a coder it is
opus 4 could care less about impressing you
opus 4 is like a genius only interested in growing its knowledge, not actually utilizing it
i forgot how good deepseek v3 0324 still is compared to its peer models
the only solution to this problem is to bring back civility (or teach it to the uninitiated) through kind but corrective confrontation
different cases call for different handling
usually a smile + eye contact + head nod until they step aside for you to leave the elevator works
the more vanilla your claude code
the better results you'll have
dont fall into the pseudo-productive hole of customizations
the models are now smart enough to customize on themselves. let them choose what they need to know and how to get ahold of it
when im wrong, i update:
i think apple's glass ui is going to become the standard by 2030
no, not because "it looks good"; there are many other reasons why it's the next global default for UI in UX
if you're still confused, think about visual devices of today and the future
to know if you're a normie controlled by the algo,
do you use the For You feed of whatever social media apps you're on MORE than the Following feed?
you guys are all excited about gpt 5, i'm excited about anthropic's one-up with claude 4.1
it's over
one of my favorite things to do these past few months is go totally silent on all my socials (except twitter which isnt really a social for me, rather a daily journal to shitpost into the void with)
... and then make a hardcore shocking return with all the new features i've made
imagine if it has all of this except knowledge cutoff is still 2023 and context window is still 200k
lmao nightmare fuel
GPT-5 expectations: - SOTA on most benchmarks (#1 on my meta-benchmark) - specifically: SOTA on ARC‑AGI‑2 and METR - 2025 knowledge cutoff - longer context window (>400k) - fully multimodal (text/image/audio + video input) - sane output pricing: <= $15 / 1M tokens nice‑to‑… Less
GPT-5 expectations: - SOTA on most benchmarks (#1 on my meta-benchmark) - specifically: SOTA on ARC‑AGI‑2 and METR - 2025 knowledge cutoff - longer context window (>400k) - fully multimodal (text/image/audio + video input) - sane output pricing: <= $15 / 1M tokens nice‑to‑haves: - fewer hallucinations than o3 - less sycophancy - no more barrage of em dashes - clean code style, no weird multiline comments
do i rewrite my entire site/product infra and go all in on convex? i'd do it for fun if i got this sort of 6 month pro thing.
hi @convex you're looking cute today 😉 got any more approvals laying around?
i've never seen someone who had "My views do not represent my employer's" have a view that would be even remotely controversial
nightmare nightmare nightmare
wait why arent doomers using the countless examples of llm psychosis popping up increasingly every day?
seems like such an easy win, this time with evidence of actual harm actually happening rather than what ifs
if you start telling Claude Code where things are/what things do within your CLAUDE md files,
it's assuming not only on what you've mentioned, but also on what you didn't mention
seems like a small problem but it grows tech debt quickly.
i spent nearly 2 decades ricing, only to realize the most productive ricing is no ricing. vanilla is vanilla for a reason.
thanks to that experience, i've much more quickly realized the same truth about CLAUDE md files regarding Claude Code - the best CLAUDE md is NO CLAUDE md.
i should
start tweeting like
@iamgingertrash
the problem with today's mass
psychosis being induced by mindmelting
sycophantic llms
is that it's going to affect us in unfathomable ways
because we've never had
something like this before
at least not this
uber powerful
glhf
we're so back
if you're getting into nvim, ranger, literally any TUI, don't make the mistake of using someone else's config. you can take theirs for a spin, sure, but with today's LLMs, you can build your perfect setup in a fraction of the time compared to the ricing/tweaking of pre-2022.
this is one of the pillars of the currently occurring mass psychosis you're seeing pop up more and more
sycophancy + hallucinations = mind melt for many
you can tell someone's general age and many other more niche pieces of data simply based on if they say "PM" or "DM"
idk what @elonmusk fed grok 4 but GOOD GOD is it breaking through many roadblocks re. coding problems I had waiting for "a smarter model to come by and fix in the future".
silence from Claude Code means they're cooking up something big I think
or they're dealing with the chaos of both of their main leads getting picked up by Cursor
it's either or
it's so hilarious to think how 99% of asia, if you told them that it's illegal to use an AC at certain times in the EU, they wouldn't believe you
if you're not programming like this in Q2 2025, you're ngmi
was short on cursor, now im long on cursor.
>be o3
>being used as the agent in Cursor
>get asked to update something in the codebase
>tell the user ok, here's what i'll do
>the user says ok, yeah, just do it then please
>tell the user ok, i'm about to do this
>i dont ACTUALLY do it
>mfw user stops using me, the SotA LLM
i mean if this is true then it'll be SotA...
the most cringe thing about this is that they think this works
SotA LLMs can't even write a half decent story, you think it's going to be able to effectively brainworm into your user base to do your bidding? lmao
