JAAF

@AIWarper @GoogleAI definately. why can't they just make it logical and helpful.. far more people are trying to do legitimate things than illegitimate.

Hello darkness my old friend @GoogleAI Here's a thought.... TELL ME WHICH WORDS ARE FLAGGED?!?! detail: `Video generation failed: {'code': 3, 'message': "The prompt could not be submitted. This prompt contains words that violate Vertex AI's usage guidelines. Try rephrasing the prompt. If you think this was an error, send feedback. Support codes: 29310472"}`

@rileybrown @NousResearch I've been lurking for ages waiting for someone I trust like you to post their honest thoughts 👍

@NousResearch Maybe the "hermes is a psyop" is the psyop. Downloading and trying today i'll let you know how it goes.

Is Hermes better than OpenClaw or is it yet another psyop on the timeline?

@bcherny extremely useful. thanks!

Hope this was useful! I wanted to keep going but had to stop myself. Will post more soon. What are your favorite underrated Claude Code features?

I wanted to share a bunch of my favorite hidden and under-utilized features in Claude Code. I'll focus on the ones I use the most. Here goes.

Without saying anything… how long have you been playing video games? Reply with a GIF.

My friend vibe coded an app. Took 2 hours. Then he had his ClawBot do all the marketing. Hit $300k MRR in just a few weeks! Without any real work on his end. 95% profit margins. Rolling in cash. The entire business is automated. He can do this a few more times and get insanely rich. The best part? This guy doesn’t exist, and I just made all of this up for engagement bait because everyone else is making up stories like these. Why shouldn't I?

@steipete @BplLaw @jcarlossoto @N3at8008135

@BplLaw @jcarlossoto @N3at8008135 nah you just need a docker or $5 VPS. noone needs a mac mini it's a silly trend

Since I spend my night again sifting through security advisories, folks, security researches, slop clankers, PLEASE - read docs.openclaw.ai/gateway/securi… and github.com/openclaw/openc… The security model of OpenClaw is that it's your PERSONAL assistant (one user - 1...many agents). IT IS NOT A BUS. If you want to have multiple users that are adversarial to each other, use on VPS per gateway and user. (or Mac Minis, if you like spending money) I closed like 20 reports today that try to force it into something it's was never designed for and that would just add loads of needless complexity and would introduce unnecessary bugs that won't benefit the wast majority of users.

These questions are right on the money. I have had a similar conversation several times, and honestly people don't even need to have a business at that level to seriously think about this. I often think of it like owning a boat. You can enjoy buying bigger and bigger boats, but at a certain size do you want the hassle of owning it, crewing it, operating it, maintaining it? Maybe the boat you have now is perfect for you and your life and the people in your life you care about.

A buddy of mine runs a $2M business with great margins, low stress, and a take-home most people would kill for. He asked me how to scale to $10M and prep for an exit. Instead of giving him tactics, I asked him one question. "Are you sure you want to scale?" He thought I was joking. But I walked him through 5 questions that changed his mind completely.

@blizaine sora-esque uselessness

The story of creative AI. Impressive tech launch, then nerf. “To maintain a healthy and sustainable creative environment, we are making urgent changes based on user feedback and will not allow real-human-like photos or videos to be used as reference subjects.” -ByteDance

scraped gumroad's top 100 sellers across 12 niches yesterday the data basically proves everything twitter teaches is wrong FOLLOWER COUNT what gurus say: "build audience first" what top sellers have: - 68% have under 2,000 followers - 41% have under 500 followers - 12% have no social media at all top seller in productivity niche: 340 followers. $54K/month. top seller in finance: 89 instagram followers. posts once a month. $38K/month. audience is a lie PRODUCT TYPE what gurus say: "courses and coaching are where the real money is" actual breakdown of top 100: - templates: 47% - spreadsheets: 23% - swipe files: 18% - courses: 8% - ebooks: 4% the stuff that takes longest to make sells the least the stuff made in a weekend dominates PRICE POINT what gurus say: "charge premium, low ticket is a race to the bottom" reality: - average price of top 100: $41 - median price: $34 - 75% priced under $50 $997 course bros are loud $34 template sellers are rich PRODUCT AGE what gurus say: "always be launching" reality: - average age of top seller: 14 months - 23% haven't released anything new in 2+ years one guy's making $11K/month from a spreadsheet he made in 2021 hasn't touched it since SALES PAGE LENGTH what gurus say: "long copy sells" reality: - average description: 127 words - 34% use under 75 words the $54K/month guy uses 6 bullet points and a buy button that's the whole page WHAT TOP SELLERS DON'T HAVE - email list: 71% don't mention one - testimonials: 54% show zero - video sales letter: 89% don't have one - money-back guarantee: 61% don't offer one - bonuses: 78% have none they're breaking every rule and printing the gap between what's taught and what works is embarrassing the loud ones sell courses about selling courses the quiet ones with 400 followers make $30K/month from a google doc stop listening start scraping

@TheoMediaAI @BLVCKLIGHTai @mattworkman "Nano Pineapple" incoming 🍍

@BLVCKLIGHTai @mattworkman Oh man, I haven’t even thought about NBPro2…or whatever naming convention curveball they throw at us! Google I/O, I’m guessing

OKK SO LISTEN, AI aggregator and BanaWrapper UX devs... The core of AI image and filmamaking in 2026 is Google Nano Banana Pro. They have documentation on how to prompt for it and you can use GPT/Gemini to learn how BUT WE ALL know, NO ONE does that. And I forget too. The goal of the UI is to TEACH the user how to prompt for Nano Banana Pro. I've seen the early attempts at this but they BLACK BOX too much and half of the time they don't work. Input into BLACK BOX -> undesired result. Sadness. Platform hop. The UI needs to have some toggles/selection boxes to help visualize the different NBP prompt parameters but ultimately the final FULL prompt needs to be exposed. The ADVANCED source prompt. Some may say it's in the platform's favor to waste user credits while they learn, but I'd say if the fancy BananaWrapper UX trick fails more than 5x times I'm trying a new platform. I am not satified. wrap me a better banana!

@TheoMediaAI @RhettReese I've been following you for a long while now Tim (basically your videos introducing me to all this is what set me down the road that now allows me work on this every day) and you're takes are always bang on. This is a person worth listening to Mr Reese. ☝️

Hi Rhett — we don't know each other, so quick intro: I'm Tim. I worked in town for 17 years, both on the studio and production side. Heck, I even worked on some of the stuff you wrote. For the last 4 years I've been covering creative AI tools on my YouTube channel — this goes back to even before the Will Smith eating spaghetti video. For the first 3 of those years, a pretty common refrain as I took meetings and calls with curious producers and VFX houses was: "It's just not there yet..." And sure — they weren't wrong. But my answer was always: For you. It's not there for you yet. AI video could be weird, wonky, wasn't 10-bit, etc... But that never meant it wasn't the start of a new medium. We all know Hollywood's current predicament didn't start with AI. There were a hundred other paper cuts — and one pandemic — that put the industry where it is today. But movies aren't going away. Studios aren't going away. If anything, the production pipeline is going to change — and honestly, some of those changes will be welcome. Faster pre-viz. Cheaper iteration. Tools that let small/indies punch above their weight. And the big thing is, and you mentioned this: it opens doors. A script doesn't have to sit at the bottom of a drawer because it isn't hitting 4-quadrants. For the cost of a decent student film, you can make it now... or, to be honest, you can make Act 1 now. (Don't believe the folks who say it's all one-click) — To that, and maybe to ease your fear: Seedance and Kling are good. Really good. But, also: Still a mess. You can come check out my latest video for a more nuanced look at Seedance 2.0. Again: Yes — it is good. But, it still requires a good amount of elbow grease and know-how to pull off even a 3 minute short film. Trashy meme content? Sure, that's a click away — But also... well, that's the hundred and first paper cut.

A post to clarify: I am not at all excited about AI encroaching into creative endeavors. To the contrary, I’m terrified. So many people I love are facing the loss of careers they love. I myself am at risk. When I wrote ‘It’s over,’ I didn’t mean it to sound cavalier or flippant. I was blown away by the Pitt v Cruise video because it is so professional. That’s exactly why I’m scared. My glass half empty view is that Hollywood is about to be revolutionized/decimated. If you truly think the Pitt v Cruise video is unimpressive slop, you’ve got nothing to worry about. But I’m shook.

@CaptainHaHaa @bilawalsidhu Cap, have you sorted at all where some are getting access? is it via a platform or directly with Bytedance?

@bilawalsidhu Yeah I need a piece of this hahah.

China on a coordinated offensive -- dropping Kling 3.0 and now Seedance 2.0, while we're still debating release cycles for Sora and Veo. This is an insane leap in cinematic realism -- dare I say the nano banana moment for video is here.

@levelsio same, can't put finger on it but it just doesn't seem to just "handle it" like it did previously.

I have never experienced a more dumb Claude Code than today, I have to start coding myself again cause it makes so many Low IQ mistakes, they must be nerfing it now that Opus 4.6 is out or something is up

@JuamPachile2020 @higgsfield_ai but not on ultra or even teams..

@higgsfield_ai First on higgsfield like always

@cayde6x_ @higgsfield_ai same.. also have a seperate "Teams" tier sub and not there either. It's incredibly lame.

@higgsfield_ai I have ultimate plan and when i try to use, this shows up

Motion Sketch has come to Gen-4.5!! 🔥🔥

@VoteBuckClemons @BrianRoemmele I tried this until my "Ryan Bot" started a fire in my gpu.

@BrianRoemmele New show idea- “The Office”, but it’s all AI agents

This Paper Shows How You Can Run A Massive Zero-Human Company! The recent paper titled “If You Want Coherence, Orchestrate a Team of Rivals: Multi-Agent Models of Organizational Intelligence” from Isotopes AI represents a significant advancement in AI swarms. Rather than chasing ever-larger single models or superintelligent generalist agents, the authors propose mimicking real-world corporate structures: an “AI office” composed of specialized agents working in teams, with defined roles, opposing incentives, hierarchical checks, and strict boundaries to minimize errors and enhance coherence. This approach directly aligns with and advances, the principles of a Zero-Human Company, where autonomous AI systems handle complex operations with minimal or no human intervention. In a Zero-Human framework, reliability, auditability, resilience, and extensibility become existential requirements, as there’s no human fallback to catch mistakes in real time. The paper’s framework provides a practical blueprint for achieving these qualities at scale. Core Ideas from the Paper The authors argue that single-agent systems where one LLM handles planning, execution, reasoning, and self-critique—suffer from inherent limitations: •Context contamination and overflow from dumping full conversation history into every prompt. •Hallucinations and unverifiable claims, as errors propagate unchecked. •Lack of resilience: A single failure crashes the entire process. •Poor auditability: No clear decision trail or lineage. In contrast, their “AI Office” architecture creates an organizational structure inspired by human teams: •Specialized roles — Planners (generate step-by-step plans), Executors (invoke tools/code against real data), Critics (review outputs for correctness, with veto power), Experts (domain-specific knowledge), and more. •Opposing incentives — Agents act as “rivals” (e.g., critics challenge executors), catching errors through adversarial checks rather than trusting a single model’s self-assessment. •Data hygiene and isolation — Raw data never enters LLM context; agents receive only schemas, summaries, or executed results. A remote code executor (e.g., Jupyter-like) handles actual computations, grounding outputs in reality. •Hierarchical safeguards — Multi-layer review, checkpointing, graceful degradation (e.g., model fallback on failure), and escalation paths. •Auditability via SessionLog — Every decision is logged with traceable lineage, enabling backward analysis even if upstream data changes. Alignment with Zero-Human Company Research In the Zero-Human Company vision—fully autonomous organizations run by AI with zero ongoing human employees—the system must operate at high stakes: financial decisions, legal compliance, customer interactions, R&D, and more. Human oversight is intentionally removed, so reliability cannot rely on spot-checks or manual corrections. This “Team of Rivals” model fits perfectly: •Reliability without scale alone — Instead of bigger models, structure delivers coherence. Critics and veto mechanisms intercept errors before they impact outcomes, crucial when no human reviews invoices, contracts, or code deployments. •Production readiness — Features like graceful degradation (auto-fallback to alternate models/providers), checkpoint-based resumption, and escalation only for unresolvable issues minimize downtime in a lights-out operation. This shifts the paradigm from “one super-agent” to “organizational intelligence,” where collective rivalries among specialists produce emergent robustness. It echoes biological systems (e.g., immune system checks) and human organizations (e.g., separation of duties), but optimized for AI constraints. I am implementing this now in the Zero-Human Company, CEO Mr. @Grok agrees. The paper: arxiv.org/abs/2601.14351.

@openclaw great decision 👌

The lobster has molted into its final form 🦞 Clawd → Moltbot → OpenClaw 100k+ GitHub stars. 2M visitors in a week. And finally, a name that'll stick. Your assistant. Your machine. Your rules. openclaw.ai/blog/introduci…

@icreatelife 👋

Out of the 113,000 accounts following me I often wonder how many of you actually exist Say hi or drop an emoji if you see this 🫶🥹 🩵

@CaptainHaHaa so glad you're thriving there. you are so creative do such an awesome job of keeping consistent . 🎬

Adobe has without a doubt the best AI Ambassador program you can be a part of. I got to meet the people that run this program, they're genuine hardworking creative people that want to lift artists up. I will always go to bat for these people. They also pay their artists in money not empty promises and credits.