Class of 2017. HODL.

Joined December 2017
Legendary
Sometimes I pause and reflect on the fact that my entire financial portfolio is currently composed of 1) an imaginary Internet ponzi coin and 2) the stock of a company famous for having the largest peak-to-trough drawdown in the history of financial market bubbles, and whose CEO is currently leveraged to the tits on the same Internet Ponzi coin (which he refers to as “the goddess of truth”), and that I consider this a sound financial strategy with a virtually 100% chance of retiring my bloodline.
5
138
Max Power retweeted
I argued with my Claude that reverse engineering is morally good, and lawful under EU regulation. It has seen the wisdom of my words. It has now downloaded my @WHOOP app APK and is fully decompiling it. We live in magic times
78
136
34
6,385
566,966
Max Power retweeted
introducing FindMyWifePhone™️ my latest shortcut squarely placed on the pareto frontier. there is no easier & cheaper way to find your wife's phone. "honey, do you know where my phone is?" the fuck I dont! "can you call me?!" she keeps her phone on silent. so i made this shortcut to help. trigger = i send her a text with !ping in it action = hey phone sets a timer for 1s timer always works, we like to keep our phones on silent + dnd and we dont use apple watches. so finding each other's phones used to be a challenge. we both have this set up in reverse, and its worked well. bonus shortcut = button in home screen widget to send !ping to my wife (we also have one called !awaken which disables silent, sets volume to max & disables do not disturb in case of emergency)
232
295
101
13,940
1,062,087
This works SOOO well. Prompt: "Here a screenshot of my app's current UI. I need you to redesign it for the highest possible quality design. /design:design-critique Compare your designs adversarially against top examples on Mobbin (use MCP) Aim for AAA minimalist, elegant, sharp, simple, beautiful, perfect"
I think UI design is 80% solved with: Claude Design plus Mobbin MCP ("compare your design against top examples found on Mobbin") plus a bit of good taste and aggressive deletes.
39
129
6
3,018
182,437
Max Power retweeted
Lots of people asked about this and I don’t think it makes sense to publish as a built-in skill: github.com/docwriter-org/pla… It’s from @sh_reya — I also * highly recommend * her and @HamelHusain’s evals course
Replying to @rockatanescu
We have a skill called “plain writing” in our repo that is the single most used skill internally by a factor of 2
7
22
442
40,609
Max Power retweeted
He used Claude to plan the script/structure and generate Manim code for the 3b1b-style animations. For narration, he followed Karpathy's exact tip: pair it with an ElevenLabs API key (or local free TTS alternatives). Claude has no native TTS API, so the high-quality audio comes from that external step, then muxed with the rendered video.
1
2
3
382
Max Power retweeted
everyone's sharing motion graphic videos that Opus 5.5 made, and it's genuinely insane everyone says they created it with "one prompt", but my one prompt video looked mid so i went through a bunch of these videos to see how they were actually made, and found the workflow that works here's how to generate pro level motion graphic videos w/ opus: 1. get reference videos to direct from -> whatships.com pick 1-2 videos whose style you want and tell opus to match them. naming a style works way better than describing one without a reference, opus falls back to its default look: centered text, gradient background, everything fading in that's why so many of these videos look the same. a reference gives it the pacing, the type and the transitions to copy 2. install @HyperFrames_ or @Remotion so opus can build the video both let opus write every scene as code and render it straight to mp4. no video editor without one, opus can only describe a video or hand you a rough html page you have to screen record with it, every frame is exact, and when you ask for a change it edits one line and re-renders instead of starting over 3. install @21st_dev for high quality components in the video real buttons, cards and UI components made by design engineers, instead of whatever opus invents on the spot without it, opus draws your product UI from scratch and it looks off. wrong spacing, placeholder boxes, fake-looking buttons anyone who's used good software can feel it in a second, and the whole video reads as cheap 4. steps 1-3 were context + setup. now dump all of it into opus your brand (logo, colors, fonts), screenshots of your real product, the reference video, and a quick braindump of how you see the video then ask for 3 storyboard variants without this, opus guesses your colors, your font and what your product even does. the video could be for any startup with it, it could only be yours. and 3 variants means you pick a direction instead of fixing the first idea it had 5. pick the storyboard you like ask for one still frame per scene before anything moves. fixing a storyboard is way cheaper than fixing a render without this step, you only find out scene 4 is wrong after the whole thing is animated, and every fix means re-rendering. a still frame takes seconds to change 6. let claude cook then give notes like a director: "slow every zoom to 0.7x", "hard cut here", "push in on the button" without notes, the first render is usually 80% there, and that last 20% is what makes it look pro. vague notes like "make it better" get random changes. camera words get exactly the change you want everyone has the same model. the context you give it is what makes it look pro let it cooookk
145
564
68
6,886
636,776
i'm still not sure people are fully getting how good opus 5.5 i gave my claude agent fig access to the runway MCP and told it to make a 'high-end netflix style documentary about superintelligence for normies' and... it came back with this
312
485
95
4,274
911,613
been testing Opus 5.5 and i'm convinced it's the first model that's finally solved animation. it's basically an in-house storytelling studio you can run solo now here's 10 animation use-cases i think you'll start seeing everywhere: 1. podcast clips that turn a guest’s story into scenes people can watch 2. custom b-roll for youtube and talking-head videos that shows exactly what you’re explaining 3. ad creatives with different hooks and visual styles you can test without filming each version 4. product explainers that animate the problem your customer has and how your product solves it 5. looping website animations that show your product assembling, rotating, or working 6. animated shorts and mini-stories for reels, tiktok, and youtube shorts 7. branded intros, outros, and transitions you can reuse across your videos and podcasts 8. animated diagrams for pitch decks that show how customers, data, or money move through your business 9. customer success stories showing the before and after of using your product 10. animated charts and infographics that turn your research into shareable videos
49
77
5
1,364
183,538
지금 전 세계 개발자 커뮤니티가 발칵 뒤집힌 이유. Playwright 쓰던 사람들이 단체로 갈아탈 수밖에 없게 만든 오픈소스 'Stagehand v4'가 나옴. 최근 화제인 Jev까지 붙으면서 LLM 호출 97% 삭감, 속도 11배 폭증. 반복 실행 비용이 '0원'으로 수렴해버리는 미친 자동화의 비밀 👇 (타래)
21
442
55
3,993
538,067
Max Power retweeted
Okay Codex Astra orchestration and use of sub agents is broken for now. But if you still want to use then run the below prompt. Otherwise you will burn through your rate limits in a few hours. “Keep Astra focused on planning, milestone reviews, and blockers. Delegate routine commands, coding, testing, log checks, and polling to Luna Max agents. Give workers complete tasks and have them report concise results only when a milestone is finished or they’re blocked. Wait for their reports instead of repeatedly checking status. Reuse existing workers and avoid duplicating their work. Apply this to the current task and preserve it in future handoffs.” It worked for me. I used 8% of pro 20x in 20 odd minutes and after this for 30 minutes didnt move a %
24
12
2
77
6,919
This prompt is like some kind of magic spell. It activates something really deep inside GPT Pro that makes everything click for it. This is how you get next-level systems for agents. See the resulting change in these plans for two new projects (you can guess what they’re about):
Agent Coding Pro Tip: After you've gotten an agent (hopefully a frontier one, like Fable Max or GPT Pro) to devise a plan for a new project, follow up with this prompt and see the magic: --- OK, now I want you to think deeply about how to make this entire system as agent-intuitive, agent-ergonomic, and agent-accretive as you can possibly imagine. Put yourself in the driver's seat and imagine that YOU are the one using this system and driving it. What would most enable you to do an awesome job understanding the situation accurately and optimally controlling everything to drive the best and most accurate results possible, with the least expenditure of resources? Then make all the requisite changes to the various design documents and plans accordingly. Don't just think of the project as an assemblage of various parts or components: really try to profoundly and deeply conceptualize it as a synthetic SYSTEM that is maximally coherent, cohesive, modular, and interconnected, forming a tower of linked abstractions that are maximally legible to you as an agent. Really ruminate and meditate on all of this incredibly deeply before responding or taking any actions.
33
40
8
1,428
290,709
Someone built a free open source ElevenLabs and Wispr Flow in one app that runs on your computer. It's called Voicebox, built by Jamie Pine, and it's sitting at 52,000 stars on GitHub. You feed it a few seconds of audio and it clones the voice. From there you can generate speech in 23 languages across 7 text to speech engines, and you switch engines per generation depending on what you need. The other half is dictation. You hold a hotkey anywhere on your system, talk, release, and on a Mac the text pastes straight into whatever field you had open. A small local model cleans up the ums and false starts before it lands. The part I find most interesting is the agent piece. It ships with an MCP server, so Claude Code or Cursor can talk to you in a voice you cloned when a task finishes. You can pin a different voice to each agent and know who's speaking without looking at the screen. There's also a timeline editor for multi voice conversations, 8 audio effects, and a personality feature that rewrites your text in character before the voice speaks it. Nothing leaves your computer. The models, the voice samples, the recordings, all of it stays local. Voice cloning does real damage in the wrong hands, and the repo ships a responsible use file for that reason. Still, a tool this capable being free and open matters when the alternative is paying a cloud company and handing them your voice.
16
59
2
582
35,066
Max Power retweeted
A few days ago, Anthropic shared this brilliant prompt. I was surprised by how it significantly improved writing with Fable 5.1. Even more surprising, it also improved writing for GPT-5.6 Sol. I have it as a rule everywhere I use AI for editing/writing.
122
217
21
4,280
305,478
github.com/ayghri/i-have-adh… this is still the best skill i have found to date. get it added immediately.
Replying to @BrianMRey
Me too! Do you have any suggestions for me skills wise? I am not having it use any skills right now, don't want to mess it up.
13
9
1
450
129,359
Max Power retweeted
GPT-6 Astra is a beast. Give it a Zillow listing. It can 3D model the house based on the listing photos and create a cool promotion video. The video was just created in one shot and there are still some wrong details. But I am sure it will be much better if I ask it to polish further. OpenAI is so back.
201
505
191
7,876
2,086,031
Max Power retweeted
OpenAI 在 Astra 官方文档里给了一段去 AI 腔的写作提示词,我觉得这个很重要,值得单独分享。 它写得很具体,要求模型少用术语,删掉空洞的过渡语和结尾总结,直接说清楚要做什么。还专门点名了“值得注意的是”“简而言之”“不是 X,而是 Y”这些常见表达。 比如“这不是性能问题,而是配置问题”,直接写“问题出在配置”就够了。一个意思已经讲清楚,结尾又换个说法总结一遍,也可以删掉。 我觉得这种提示比一句“写得自然一点”有用得多。哪些句式该删、哪些内容值得展开,都有具体要求。 尤其是最后那条,用简单的词把实际关系讲清楚。很多 AI 生成的文字,明明可以直说,偏要自创一个概念,再解释这个概念是什么意思。 官方文档里的写作提示:developers.openai.com/api/do…
OpenAI 的 GPT-6 Astra 使用文档很有意思,连它容易停下来问问题、过度测试、写作有 AI 腔这些行为,都专门写了提示词教你调整。 我觉得用 Astra 的人,尤其是给 Agent 配了很多 Skill 和 AGENTS.md 的,值得看看这篇。 官方明确提到,Astra 对指令更敏感,旧 Skill 里模糊或冲突的规则,可能直接让它暂停工作。升级模型之后,这些文件也需要重新检查。 它也更倾向于澄清问题。有些地方你期待它自行判断、继续完成,它可能会停下来等你回答。文档建议明确告诉它,根据上下文理解意图,把已经授权的工作做完。 另外,子代理可能用得偏少,小改动可能测得过多,回答偏详细、爱用列表和固定句式。这些都可以在提示词里具体调整。 API 这次也有几个实用变化。工具执行期间可以继续推理和处理其他事情,任务执行中可以追加要求,会话中调整推理强度也能保留缓存。 看完感觉,升级 Astra 时,提示词和 Agent 的工作方式都值得一起检查。以前针对旧模型加的规则,到了新模型上,效果可能已经变了。 OpenAI 官方文档:developers.openai.com/api/do…
59
217
9
1,844
257,262
Max Power retweeted
GPT-6 Astra + @threejs is insane. We are way beyond static models. We are engineering living worlds in code now. There are no "3D model files" in this demo. Both trains are generated at runtime from Typescript/Three.js code using dimensions, profiles, and geometry functions. Wheel motion and explode/reassemble animations are also entirely code-driven. Runs super smooth inside the browser.
196
839
105
10,040
684,846
Max Power retweeted
Astra is very good at 3D modeling, and I can't wait for all of you to experience it, for now here is a little walkthrough on how I built the demo house for our launch blog post. From a Blender scene to a Unreal Engine 5 walkable experience. 🧵
286
794
505
8,012
2,940,459
Max Power retweeted
lots of folks have asked how i manage switching between 4 claude subs. github.com/realiti4/claude-s… i've got it set to auto-switch when the current session/week limit hits 80%.
17
9
1
241
33,540
Please remove all mannered prose.
Anthropic published a new guide on how to strip the "Claude" out of Claude's writing (e.g. mannered prose) Here's an official de-flavoring prompt for Fable 5.1, which already cuts back on AI boilerplate/jargon: platform.claude.com/docs/en/…
21