DeeKey @DeeKeyPro

The first thing YouTube shows you now might not be a video. Ask YouTube is rolling out on mobile in the US. You tap “Ask” in the search bar and type a question the way you’d say it out loud. You get a written guide back, with the videos underneath. Getting your video into that answer is Generative Engine Optimization, applied to video. I’d call it Video GEO. It changes the job for anyone who makes videos... starting with metadata: 1️⃣ write the title as the question people ask. 2️⃣ answer it in the first two lines of the description. 3️⃣ name your chapters by what happens in them, not “Part 2.” 4️⃣ fix your captions. What you say on camera is now the text that gets quoted. Make the answer easy to find. Make the video worth watching anyway. 🚀

Neal Mohan @nealmohan

2 weeks ago

Exciting news: Ask @YouTube on search is now rolling out on mobile in the U.S. 📲 Tap “Ask” in the search bar and use conversational questions to get you to the right video faster, for example: 🍂 Which national parks are best for fall colors? 🚗 What’s the best route for a 3

58 31 227 40K 44

The three most mysterious disappearances ever: • Ships in the Bermuda Triangle • Matter past a black hole’s event horizon ... • Every job application you’ve ever submitted!

Samsung just estimated Q3 operating profit up nearly ninefold, to about $80 billion. Sales up 126% year on year. The share price has more than doubled this year. But the stock still fell about 2.5% on the day. Along with $MU, $TSM, $AMD and $INTC, which are red today too. My read... when everyone already believes the AI story, the price already includes a great quarter. A record result stops being news. It has to beat what investors already bet on. I see the same thing with AI in product work: once AI is in the pitch, "it works" stops counting as a win. People want to see it in a number the CFO already tracks. 💡 So set expectations honestly before you report results. If the hype sets the bar, even a record quarter can end in a red day Source: ft.com/content/894378…

In the AI video data I've seen, longer videos came out cheaper per second. In the largest dataset, one second of finished video cost $0.155 (about $9.30/min), but that's the average! 💸 Month to month, cost per second swung from $0.092 to $0.270, nearly 3x. In the worst month, spend stayed about flat against the month before while video output fell 43%. The next month went the other way. Average video length jumped from 38 seconds to 105, and cost per second dropped to $0.098! The mechanics are simple: each session pays for planning calls, storyboard images, retries and discarded takes before a single keeper frame exists. On a 15-second clip all of that lands on 15 seconds. On a two-minute video it spreads thin. 💡 It's tempting to treat AI cost as a procurement problem: find the cheaper model, negotiate the rate... In this data the bigger levers were product calls: what format you ship and how your production pipeline is built. Plan your budget around those first!

It’s gonna be 🔥🔥🔥 @AFTERSHOCKSAC

@dshukertjr Congrats! Very well done, team 💪

YouTube changed several things this year: ▪️ Made For Kids policies got stricter ▪️ Auto-dubbing went wide ▪️ Shopping, memberships and brand deals reshaped creator revenue ▪️ YouTube started buying formats outright ▪️ The TV became the main screen ▪️ Even the definition of a 'view' moved All of it changes creator economics. None of it changes why someone picks your video over the next one. Still, the question I hear every week is: how to perform on YouTube consistently. There are answers circulating, and some of them work: a quirk in distribution, a format the system over-rewards for a while, a loophole somebody found and shared. They're real, but they're also temporary, and these windows close faster every year. So it's a choice: dig for the next lucky break, or build something that performs without one. The advice on the other side isn't better: play it safe, don't risk channel's revenue for creative pivots, control production cost, protect the margin, delegate, automate, know who owns what... That's how you keep a channel alive. It's not how you grow one. What moves a YouTube channel is smaller and harder: 🔅 Knowing your audience. Who is this for, and why would they pick it over the next thumbnail? 🔅 An idea you'd want to make even if the algorithm didn't exist 🔅 Knowing where the idea runs out. Almost any concept gives you 10 or 15 videos. Can you see year two? And is there IP at the end of it, or just more episodes? @derraleves put it well: "A system for understanding your audience, making better decisions, learning faster, and building something that can survive the next wave of change". The platform will keep rewriting the rules. A system approach and your courage are the only things that survives those changes. Research -> Make -> Analyse -> Repeat 🚀

Derral Eves @derraleves

2 months ago

👀 V2 - Coming out next week….

28 16 199 12K 48

AI can make you believe every question has a simple, one-line answer 😎 Don’t get me wrong! You can get that answer for sure... but its value will just be close to zero. Don’t stop thinking for yourself. AI is a strong first pass that saves time. You’re still the decision-maker. So be it! Here is how the top 1% of founders and operators actually use AI: ⭐️ AI is the Analyst: It handles the grunt work. Pre-analysis, synthesis, data parsing, and saving you 10 hours of research ⭐️ You are the Executive: You bring the context, intuition, edge, and ultimate responsibility AI gives you the baseline, not the breakthrough. If you outsource your final judgment to a prompt, you aren't leveraging AI... You're just turning off your brain 🤷‍♂️ Speed without critical thinking is just fast failure. Use AI to do the legwork, but own the decision!

@benjitaylor That's huge, thanks! Switching between X ads, X feed and Claude was not a bi deal, but still something I'd love to get rid off in my daily routine

@ClaudeDevs I’ve almost switched to Cursor already… You better make it permanent or I’m off

The best part of my job? Getting to collaborate with brilliant minds, partners, and specialists across our teams every single day. I’ve always relied on strong teammates. Some were recruited directly from the market, others grew into their current roles through constant shared effort, and many have moved on to build incredible careers of their own. It’s my privilege to watch their success live! Real leadership isn’t about making every call from the top, it’s about guiding and developing people, opening new horizons, and ensuring everyone understands the shared goals. True development happens when you help them see where a decision or suggestion doesn't align with those high-level goals. That feedback loop is how they sharpen their judgment and improve. The Ground Rules for Autonomous Teams 🔳 Hire domain superiority: if you’re the smartest person in the room on every single topic, you built the wrong team. Hire specialists whose depth embarrasses your own 🔳 Set guardrails, not steps: give your team high-level context, success metrics, and boundaries. Then step back and let them decide how to achieve it 🔳 Remove decision friction: enable people to make 2-way door decisions (easily reversible calls) instantly, reserving management syncs strictly for 1-way door risks When your team wins, pass all the credit down. When they fail, shield them completely. Owning the downside and backing your people is the entire point of management. To be honest, I’ve broken every single one of these rules at different stages of my growth. But learning from those missteps, and the trust built along the way, is what makes my teammates and me top achievers in our industry today

OMG! This is happening!

keyboard_arrow_left Previous keyboard_arrow_right Next

@Google Wow! The pricing wars has began!

@kwakhare5 @jacalulu @GeminiApp @antigravity Support the topic! Gemini consumption in Antigravity is barbaric, even on a paid plan 😭

@jacalulu Congrats, Jaclyn! The photo is amazing! Wanna make the same on my next trip 😍

I've been a Claude loyalist for a year. Went to Cursor to try Composer, and ended up on Grok 4.5 instead, because Cursor kept nudging me toward it 😄 Here's what my August commits (MTD) actually look like: ▪️Opus 5 — 37.3% ▪️Composer — 36.7% ▪️Grok 4.5 — 19.3% ▪️Sonnet 5 — 6.6% 🎨 Whenever I’m working on complex frontend architecture, UI polish, or component design, nothing beats @claudeai's spatial awareness and design taste. ⚡️ @cursor_ai models accounted for 56.1% of total line volume. On average, Cursor commits were ~2.4× larger—perfect for heavy-lifting boilerplate and sweeping codebase edits. With yesterday's @grok 4.6 drop (cursor.com/blog/grok-4-6) at $2/M input & $6/M output tokens, the value offer is unreal. Can’t wait to check my stats in a month!

@1Password @1Password Btw, a little link over here would help your users a lot to understand clearly: - why do I see it - how can I get rid of it (if I didn't like it) That would be a way better UX 🙏

@1Password, lease stop adding speedbumps that get in the way of daily use cases 🙏 Great UX and great safety can coexist!

@1Password > You can opt out by right-clicking the 1Password icon, selecting Settings → Notifications, then turning off Warn about potential phishing Thank you so much, team! Much appreciated 👍

@jacalulu That’s the hardest part of the process, actually 😄