"From 2003-2023, American adults reduced socializing by 30%. For teens, the decline was nearly 50%. There is no record of any period in history when people spent more time alone."
https://x.com/zarathustra5150/status/2022860794451394942?s=20
https://x.com/zarathustra5150/status/2022860794451394942?s=20
😢1
your brain literally holds on to unfinished business. this is called the zeigarnik effect.
https://youtube.com/post/UgkxgeVm1PC8Qfvss7-oflvLUS1jnI4OU4_b?si=vWPftc-Im8sdCfI8
https://youtube.com/post/UgkxgeVm1PC8Qfvss7-oflvLUS1jnI4OU4_b?si=vWPftc-Im8sdCfI8
if the world moves towards optimization and people are beginning to ai summarize articles instead of reading through them, then in that case would the effort to result ratio of manually writing articles decrease? in that case would it be better to just generate ai articles? would this simply create a death loop or negative loop in society when it comes to creative writing?
> spend 1 week writing a 2 pager article on a topic
> readers ai summarized it and read them in 10 sec
> affiliate links, ads, expressions, etc. turn redundant
> disincentivised to write
> ai generate future articles
> readers ai summarized ...
repeat loop
i've been noticing articles talking about optimizing for ai agents and not humans anymore, when it comes to websites and anything on the internet.
SEO optimization and appearing on the first page of google used to be the game. creating hooks to capture readers attention in the first paragraph. will all these still matter when people are just summarising with AI? will these still matter when people are no longer manually searching on google, but simply asking their agents?
articles should then optimize to be picked up by AI agents instead of human engagement. being on the first page of google doesn't matter anymore, because agents can scrape thousands in seconds
i wonder how the shift in tonality, writing, and design on the internet will look in the future
my prediction is that it all converges (yes, because of ai)
> spend 1 week writing a 2 pager article on a topic
> readers ai summarized it and read them in 10 sec
> affiliate links, ads, expressions, etc. turn redundant
> disincentivised to write
> ai generate future articles
> readers ai summarized ...
repeat loop
i've been noticing articles talking about optimizing for ai agents and not humans anymore, when it comes to websites and anything on the internet.
SEO optimization and appearing on the first page of google used to be the game. creating hooks to capture readers attention in the first paragraph. will all these still matter when people are just summarising with AI? will these still matter when people are no longer manually searching on google, but simply asking their agents?
articles should then optimize to be picked up by AI agents instead of human engagement. being on the first page of google doesn't matter anymore, because agents can scrape thousands in seconds
i wonder how the shift in tonality, writing, and design on the internet will look in the future
my prediction is that it all converges (yes, because of ai)
👍3
prompt repetition improves non-reasoning LLMs
this feels like a huge tip #BM
https://x.com/burkov/status/2023822767284490263?s=46&t=nluSw7cwYGx8zFxfoYi5Iw
this feels like a huge tip #BM
https://x.com/burkov/status/2023822767284490263?s=46&t=nluSw7cwYGx8zFxfoYi5Iw
X (formerly Twitter)
BURKOV (@burkov) on X
LLMs process text from left to right — each token can only look back at what came before it, never forward. This means that when you write a long prompt with context at the beginning and a question at the end, the model answers the question having "seen"…
✍1
study the transition back to dumb devices
https://x.com/politicalmath/status/2023614697061110041?s=46&t=nluSw7cwYGx8zFxfoYi5Iw
https://x.com/politicalmath/status/2023614697061110041?s=46&t=nluSw7cwYGx8zFxfoYi5Iw
X (formerly Twitter)
PoIiMath (@politicalmath) on X
I'm so frustrated at being a parent in this garbage digital age.
I hate short-form video. TikTok / YouTube trash is just pure brain-rot and I can't stand it. I try to keep my kids away from it.
I gave them access to Spotify b/c I want them to enjoy music…
I hate short-form video. TikTok / YouTube trash is just pure brain-rot and I can't stand it. I try to keep my kids away from it.
I gave them access to Spotify b/c I want them to enjoy music…
humblespace
if the world moves towards optimization and people are beginning to ai summarize articles instead of reading through them, then in that case would the effort to result ratio of manually writing articles decrease? in that case would it be better to just generate…
X (formerly Twitter)
Nassim Nicholas Taleb (@nntaleb) on X
Until photography, painters competed largely on the verisimilitude of representation. Photography destroyed art by making artists switch to absurd images often driven by social contagion.
Now imagine what AI will do to literature.
Now imagine what AI will do to literature.
humblespace
if the world moves towards optimization and people are beginning to ai summarize articles instead of reading through them, then in that case would the effort to result ratio of manually writing articles decrease? in that case would it be better to just generate…
X (formerly Twitter)
Bryan Brinkman (@bryanbrinkman) on X
We’ve hit this odd moment where folks are using AI to write book-length articles and then viewers are using ai to reduce it back down to the size of a tweet.
😁1
once a wrong response is generated, revise the prompt entirely instead of correcting your AI down the conversation
https://x.com/hasantoxr/status/2024238760674959492?s=46&t=nluSw7cwYGx8zFxfoYi5Iw
https://x.com/hasantoxr/status/2024238760674959492?s=46&t=nluSw7cwYGx8zFxfoYi5Iw
X (formerly Twitter)
Hasan Toor (@hasantoxr) on X
🚨BREAKING: Microsoft Research + Salesforce just dropped a paper that should scare every AI builder.
They tested 15 top LLMs GPT-4.1, Gemini 2.5 Pro, Claude 3.7 Sonnet, o3, DeepSeek R1, Llama 4 across 200,000+ simulated conversations.
Single-turn prompt:…
They tested 15 top LLMs GPT-4.1, Gemini 2.5 Pro, Claude 3.7 Sonnet, o3, DeepSeek R1, Llama 4 across 200,000+ simulated conversations.
Single-turn prompt:…