•twitter5070
listening to
timeline
5070
mentions
tracked
tracked
net sentiment
+24%
1586 positive
3109 neutral
375 negative
volume & sentiment over time
Sep 30 – Oct 5, 2026 · Daily
sources
most discussed
—
active days
6
mentions
Showing 476–500 of 1586 mentions
-
• twitter 1d positivehttps://t.co/AzBuIW8eDq the kind of portfolio I’d love to have someday. absolute legend behind so much of what makes @zerodha what it is. https://t.co/AzBuIW8eDq the kind of portfolio I’d love to have someday. absolute legend behind so much of what makes @zerodha what it is.view source ↗
-
• twitter 1d positiveIt looks unbelievably simple. It was unbelievably complicated. A while back, Jon asked, "Why couldn't we just build a tool that people can use to refer people?" So I built one. You put your name in. You put your friend's name in. It finds their photo and drops you both into a retro video-game invitation. Player … It looks unbelievably simple. It was unbelievably complicated. A while back, Jon asked, "Why couldn't we just build a tool that people can use to refer people?" So I built one. You put your name in. You put your friend's name in. It finds their photo and drops you both into a retro video-game invitation. Player 1. Player 2. Every click plays a little 8-bit sound. When I showed my coach Angus, he said those sounds are "kind of like dopamine hits." Then Sandra Newson left a comment: let every attendee add one personal sentence to their invitation. So now you can add a personal note. Thank you, Sandra. I made the art and the little avatars in ChatGPT. I knocked out the photo holes in Photoshop. I started coding in Codex and moved over to Claude Code. About five hours. But the line that stuck with me was Angus's: "It becomes a contagion. If somebody gets this, they want to be a part of it." That's the whole idea behind getting to 3,000. The people already coming invite the people they'd want next to them. Want to try it live? Tell me in the comments and I'll send you the link. 👇 Press Start: The Road to 3,000. #SocialMediaExaminer #SMMW27 #RoadTo3000 #AIinMarketing #EventMarketingview source ↗
-
• twitter 1d positiveMeta. Google. Microsoft. These companies have spent years engineering faster vector-search systems. A basic nearest-neighbor query computes the distance between the query and every stored vector. Its cost grows linearly with the size of the database and becomes impractical across millions of candidates. Every vecto… Meta. Google. Microsoft. These companies have spent years engineering faster vector-search systems. A basic nearest-neighbor query computes the distance between the query and every stored vector. Its cost grows linearly with the size of the database and becomes impractical across millions of candidates. Every vector index reduces this computation differently, using clustering, graph traversal, compression, or a combination of these methods. Here are five common approaches: 1. Flat index A Flat index compares the query with every stored vector, selects the K smallest distances, and returns them. It performs an exact search, so it is useful when accuracy matters and the dataset is small enough to scan. The query cost grows linearly with the number of stored vectors. 2. IVF IVF divides the vector space into clusters. Each stored vector is assigned to its nearest cluster centroid. At query time, the database finds the closest centroids and searches only their clusters. A probe parameter controls how many clusters it searches, and increasing it usually improves recall but also increases latency. 3. HNSW HNSW stores vectors in a multilayer graph. The upper layers contain fewer nodes and allow large jumps through the graph. Lower layers contain more nodes and support a finer search. A query starts at the top, moves toward closer nodes, and descends through the layers. HNSW often provides low latency and high recall, but the graph requires extra memory. 4. IVF-PQ IVF-PQ combines clustering with product quantization. IVF limits the search to a few relevant clusters. Product quantization splits each vector into smaller parts and stores compact codes for them. Comparing these codes is cheaper than comparing full-precision vectors. This reduces memory use and query cost, with some loss in recall. 5. ScaNN ScaNN partitions the dataset and stores quantized vector representations. During search, it selects relevant partitions, scores compressed candidates,view source ↗
-
• twitter 1d positiveI love SI slopcore I love SI slopcoreview source ↗
-
• twitter 1d positiveRT @ghita__ha: when the thing you built 2 years ago suddenly becomes cool again 😂 RT @ghita__ha: when the thing you built 2 years ago suddenly becomes cool again 😂view source ↗
-
• twitter 1d positiveToday x gives me ballons.. (Actually it didn't, hahah,wtf x?😝) Thank God for every day!! 40 here I come.. https://t.co/xLj42V2EZB Today x gives me ballons.. (Actually it didn't, hahah,wtf x?😝) Thank God for every day!! 40 here I come.. https://t.co/xLj42V2EZBview source ↗
-
• twitter 1d positiveThe reason designers can't build sites like this isn't talent. It's that nobody tells them step 1. Save this. Repost it so another designer finds it too. Here's how it actually works 👇 1. It's not a video. It's hundreds of still images played in order - like a flipbook. Scrolling replaces play/pause. 2. One image s… The reason designers can't build sites like this isn't talent. It's that nobody tells them step 1. Save this. Repost it so another designer finds it too. Here's how it actually works 👇 1. It's not a video. It's hundreds of still images played in order - like a flipbook. Scrolling replaces play/pause. 2. One image stays fixed on screen the whole time. Everything else just decides which frame of the flipbook you're looking at. 3. Your scroll position = one number. That single number controls everything - which image shows, how zoomed in it is, when the text changes. 4. Break the site into "scenes," like chapters. Scene 1 plays for the first bit of scrolling, then scene 2 takes over, and so on. 5. Between scenes, don't just cut - blend. A slow dissolve or wipe between images is what makes it feel expensive instead of choppy. 6. Load images smart: rough version first, sharp version once the reader scrolls close. That's why it never lags or freezes. 7. Text appears and disappears on its own scroll range too - tied to that same number, so it's never out of sync with the images. That's the whole trick. No video editor, no complicated code - just images, one number, and good timing.view source ↗
-
• twitter 1d positiveBeauty + Power in your pocket 🙂↕️ Empryo users eating good https://t.co/lZnpxY9oFI Beauty + Power in your pocket 🙂↕️ Empryo users eating good https://t.co/lZnpxY9oFIview source ↗
-
• twitter 1d positiveThings will go wrong. You will screw up. You’ll lose money, friends, opportunities. But you must never lose the fire inside you that drives you to making it. Once you develop the ‘I’ll show up again tomorrow’ mindset, you win. Things will go wrong. You will screw up. You’ll lose money, friends, opportunities. But you must never lose the fire inside you that drives you to making it. Once you develop the ‘I’ll show up again tomorrow’ mindset, you win.view source ↗
-
• twitter 1d positiveSunday is for builders. Share what you're working on + your city 👇 I'll check every single one. Sunday is for builders. Share what you're working on + your city 👇 I'll check every single one.view source ↗
-
• twitter 1d positiveWANT MORE PAGES READY TO RANK? STEAL MY 200+ AI SEO PROMPTS 📈 I packed my SEO prompts into one FREE 50+ page book. 📚 Use them for: 🎯 Keywords ✍️ Content 🗺️ Topical maps 🔗 Backlinks 📈 Optimization Want a copy? Comment YES 👇 https://t.co/i0Yh72y0zO WANT MORE PAGES READY TO RANK? STEAL MY 200+ AI SEO PROMPTS 📈 I packed my SEO prompts into one FREE 50+ page book. 📚 Use them for: 🎯 Keywords ✍️ Content 🗺️ Topical maps 🔗 Backlinks 📈 Optimization Want a copy? Comment YES 👇 https://t.co/i0Yh72y0zOview source ↗
-
• twitter 1d positivei fixed the Codex app, thoughts? https://t.co/dGZxuamE3D i fixed the Codex app, thoughts? https://t.co/dGZxuamE3Dview source ↗
-
• twitter 1d positiveThe coding agent says the tests pass. I ask which tests. It opens the terminal like we're both new here... The coding agent says the tests pass. I ask which tests. It opens the terminal like we're both new here...view source ↗
-
• twitter 1d positiveYour AI can build interfaces. These skills help it build better ones. Save this ↓ - frontend-design https://t.co/3SEZWdagEn - apple-design https://t.co/9Kt8f6xsNd - beautiful-shadows https://t.co/mzli9De6Gn - accessibility https://t.co/zPXMJni6Mq - design-review https://t.co/9DQ00nr1LB - emil-design-eng… Your AI can build interfaces. These skills help it build better ones. Save this ↓ - frontend-design https://t.co/3SEZWdagEn - apple-design https://t.co/9Kt8f6xsNd - beautiful-shadows https://t.co/mzli9De6Gn - accessibility https://t.co/zPXMJni6Mq - design-review https://t.co/9DQ00nr1LB - emil-design-eng https://t.co/wTG5OxdsKw - shadcn https://t.co/l4j6j6B8dD - adapt https://t.co/njnkLxoXzQ - better-interface https://t.co/liznEhyqGb - interaction-design https://t.co/LqDrynuPKM Which one are you adding first?view source ↗
-
• twitter 1d positiveOne of the best things in life is seeing a smile on a persons face and knowing that you put it there. One of the best things in life is seeing a smile on a persons face and knowing that you put it there.view source ↗
-
• twitter 1d positiveStart with a dream. Turn it into a goal. Break it into steps. Make those steps part of your day. Build something your future self will thank you for and then enjoy the rewards. Start with a dream. Turn it into a goal. Break it into steps. Make those steps part of your day. Build something your future self will thank you for and then enjoy the rewards.view source ↗
-
• twitter 1d positiveTo all the “I don’t check my code anymore” crowd: How would you react if you discovered that your entire flow was built on regexes and patches acting like landmines in production? Do you understand that no harness can save you from slop? ** This is a real screenshot of top tier model trying to refine a loop ** To all the “I don’t check my code anymore” crowd: How would you react if you discovered that your entire flow was built on regexes and patches acting like landmines in production? Do you understand that no harness can save you from slop? ** This is a real screenshot of top tier model trying to refine a loop **view source ↗
-
• twitter 1d positiveGood afternoon, X! ☀️ How’s your Sunday going? Working, building, or just chilling today? 👀 Good afternoon, X! ☀️ How’s your Sunday going? Working, building, or just chilling today? 👀view source ↗
-
• twitter 1d positiveThis is a big one for AI creators. 🔥 Having MCP connect GPT, Claude, or other agents directly to the InVideo editor makes the entire creative workflow much more flexible. Excited to see how far we can push this! @invideoOfficial This is a big one for AI creators. 🔥 Having MCP connect GPT, Claude, or other agents directly to the InVideo editor makes the entire creative workflow much more flexible. Excited to see how far we can push this! @invideoOfficialview source ↗
-
• twitter 1d positiveWhat a killer video I made for @WisprFlow 😍 Crazy that I can prompt this and it was made in 10 mins 😳 All thanks to the SKILL file I created reply with "motion skill" below and I'll send it to you https://t.co/YHN6RkLlkN What a killer video I made for @WisprFlow 😍 Crazy that I can prompt this and it was made in 10 mins 😳 All thanks to the SKILL file I created reply with "motion skill" below and I'll send it to you https://t.co/YHN6RkLlkNview source ↗
-
• twitter 1d positiveuse only ONE PHONE per app on tiktok/insta - get yourself an iphone SE or 11 - reset to factory default - install tiktok and insta - sign up with a new account - create email with outlook or zoho (it's free and fast) - post your videos you can find those phones easily too. start posting asap use only ONE PHONE per app on tiktok/insta - get yourself an iphone SE or 11 - reset to factory default - install tiktok and insta - sign up with a new account - create email with outlook or zoho (it's free and fast) - post your videos you can find those phones easily too. start posting asapview source ↗
-
• twitter 1d positiveopenai has been working on the research behind dots for almost two years. most people assumed this was a response to other labs. It wasn't the thing that took two years wasn't making it smart, it was making it remember. → dots stay coherent over weeks. you teach it your preferences once, it doesn't forget the brief … openai has been working on the research behind dots for almost two years. most people assumed this was a response to other labs. It wasn't the thing that took two years wasn't making it smart, it was making it remember. → dots stay coherent over weeks. you teach it your preferences once, it doesn't forget the brief two weeks later. → they never stop. if there's nothing useful to do, they sleep. the moment something comes up, they wake back up. → they run long-horizon tasks across multiple days without losing track every other agent today resets context after a conversation ends. dots don't have conversations. they have ongoing relationships with your work. that's what took two years to build.view source ↗
-
• twitter 1d positiveToday. The 16VC Accelerator application closes. $200K. No program fee. 3 months. SF · NYC · London. Still early? That's okay. If you're building, apply today. #16VC #Startups https://t.co/YLFj1bQlAD Today. The 16VC Accelerator application closes. $200K. No program fee. 3 months. SF · NYC · London. Still early? That's okay. If you're building, apply today. #16VC #Startups https://t.co/YLFj1bQlADview source ↗
-
• twitter 1d positivePOV: After you paid $5000 to learn how to build a second brain with opus 5.5 then find the FREE version below https://t.co/26AKqmSzBG POV: After you paid $5000 to learn how to build a second brain with opus 5.5 then find the FREE version below https://t.co/26AKqmSzBGview source ↗
-
• twitter 1d positive🚨 THIS IS CRAAAAZY! NotebookLM can now turn YouTube tutorials into useful notes. Here are 10 prompts to help you do it all 👇 https://t.co/M6IBdfLeAK 🚨 THIS IS CRAAAAZY! NotebookLM can now turn YouTube tutorials into useful notes. Here are 10 prompts to help you do it all 👇 https://t.co/M6IBdfLeAKview source ↗