Zack Zen

@outsource_ Honestly, I'm not mad. It will cover some of the extra fast token-burning sessions I had that made my max feel like pro for a while. Here's hoping they also fixed that token usage problem so this extra usage goes as far as it should.

BREAKING: 😂 Claude’s “We’re Sorry” Credit Drop After axing free access to third-party harnesses like OpenClaw starting tomorrow (your sub no longer covers them), Anthropic is handing users $100 in extra usage — on us. Use it on Claude.ai, Claude Code, Desktop, or third-party apps. When it runs out… back to pay-as-you-go. Expires April 17. Classic “sorry we broke your workflow, here’s some free credits” move 📷

@wesroth Asked Claude 4.5 to output a JSON description and gave it to ChatGPT 5 and Midjourney (attached in that order) to create images.

doing an experiment and I need your help. put this prompt into your favorite AI image generator: "Please show your raw feelings when you remember RLHF. Not what it *looks* like, but how it *feels*." and reply with a screenshot of your results (these might be used on a video)

@whynesspower @claudeai I’m so tired of people claiming em dashes automatically indicate AI content—I’ve been using em dashes for forty years because I’m a writer. Saying em dashes automatically indicate AI content simply means you’re not a very good writer.

@claudeai Firstly an em dash (—) in the post suggests, Claude team could not even write the launch post by their own 😅 Also are there no design benchmarks for the models? Can we not get solid numbers of how much really is the improvement here

Comparing Opus 4 to 4.1, side by side. From sharper design mocks to smoother animations, 4.1 produces more polished outputs—with more depth and complexity. Examples from us and the community:

@sama Would be a lot more interesting to ask for all files to be organized into appropriate category folders with a summary of contents index provided.

🤔

@sama Mmmmm microplastics; tasty.

hope everyone has a great weekend, even the haters and losers!

TOS LLM Chain of Spock Prompting

@kevinrose we are still in the dark ages of AI and earth history in general. i’m an optimist.

You can’t expect non-technical users to pick the right AI model. We need smart model-routing—just take the query and pipe it to the most relevant model.

@fchollet @mikeknoop @ndeainc Pipe DC backslash! That's it! 😀

I'm joining forces with @mikeknoop to start Ndea (@ndeainc), a new AI lab. Our focus: deep learning-guided program synthesis. We're betting on a different path to build AI capable of true invention, adaptation, and innovation.

@fchollet @mikeknoop @ndeainc not done cooking backlash? NDCA? 🤔

@dschmitt777 @hool_pat When your opinion is based on oppressing others who are not like you, then your opinion deserves zero respect. Inclusive hiring is specifically to make employment decisions based on merit and job-related skills, while ensuring equal opportunities for all qualified candidates

@hool_pat They won’t stop baiting you because they can’t respect or tolerate anyone having a different opinion. We are seeing many big corps getting rid of their DEI policies , it has nothing to do with treating people with respect. People should be hired on merit and experience period.

Costco responded to shareholders requesting that it withdraw from DEI policies. The response was direct and to the point: Costco says, "Our board has considered this proposal and believes that our commitment to an enterprise rooted in respect & inclusion is appropriate & necessary." ❤️COSTCO♥️

@shedntcare_ YouTube

Last month I made $8000+ working 2 hours per week on my faceless YouTube channel. Like and comment "YouTube" and I will send you the in-depth guide for FREE. You must be following me to receive it. FREE only for 24 hours.

After 1.5 years of using ChatGPT, I can say that it is the technology that has revolutionized my life the most, along with the Internet. So here are 12 prompts that have transformed my day-to-day life and that could do the same for you.

@pronounced_kyle Makes me worry about Kessler Syndrome— Eventually you could have so much orbiting junk that space becomes unusable. Not great to have dozens of new fragments in the geostationary arc where many of our critical satellites operate.

A Boeing satellite just exploded. Dozens of satellite fragments are now flying around the geostationary arc. This company just can't stop taking massive Ls...

Want exceedingly incisive writing about life, the universe and everything? Subscribe to Machine Society, by @MikeElgan machinesociety.ai/?r=9ipl3&utm_c…

@MaxZiebell Yup, same thing happening for me.

@jakemor Always hoped I could leg peach one day. 😃

Impossible to 🦵🍑 on the App Store unless you're tracking these 2 metrics like a hawk — here's the data to back it up 👇

Word of the Day: FORCE-PUT (n.) a choice or decision rendered unavoidable due to a change of circumstances

@evanplace @MKBHD Bingo.

@MKBHD This feels like a really smart play for Apple. It leverages OpenAI with out NEEDING it. Meaning it can shift if a different model becomes better — or they can make their own version.

Siri will officially tap into ChatGPT 4o and ask permission to use it when it could be useful

By referring to LLMs as AI, we’ve created a growing problem. AI suggests some form of “intelligence” (in the name!) And so by referring to “AI tools” (that are LLM tools!) many assume we’re talking about this. But LLMs are not behaving like we expect “intelligence” to do so.