•twitter15877
listening to
timeline
15877
mentions
tracked
tracked
net sentiment
+24%
4998 positive
9734 neutral
1145 negative
volume & sentiment over time
Sep 30 – Oct 14, 2026 · Daily
sources
most discussed
—
active days
15
mentions
Showing 1026–1050 of 15877 mentions
-
• twitter 2d neutralRT @jorybraun: Let's try @tokifyi, Vancouver, Canada 🇨🇦 · Nov 2nd · Science World A 3-hour workshop with @kentcdodds, followed by an afte… RT @jorybraun: Let's try @tokifyi, Vancouver, Canada 🇨🇦 · Nov 2nd · Science World A 3-hour workshop with @kentcdodds, followed by an afte…view source ↗
-
• twitter 2d neutralyour "agentic harness" will die soon https://t.co/oRoa4fzLAM your "agentic harness" will die soon https://t.co/oRoa4fzLAMview source ↗
-
• twitter 2d positiveThis morning we launched Agent Learning Services and Agent Data Services. Right now, builders at Day 0 of the Open Intelligence Summit are hands-on with open models and DigitalOcean’s agentic stack, with @SwiftAlphaOne and the @MLHacks crew leading the way. It’s the best part of open intelligence: builders choosing th… This morning we launched Agent Learning Services and Agent Data Services. Right now, builders at Day 0 of the Open Intelligence Summit are hands-on with open models and DigitalOcean’s agentic stack, with @SwiftAlphaOne and the @MLHacks crew leading the way. It’s the best part of open intelligence: builders choosing their own models and tools, and owning what they make. The summit keynote streams tomorrow at 1 PM PT. https://t.co/ujyMImUjZjview source ↗
-
• twitter 2d neutralBruh I am so done with SF https://t.co/AJLyf2EB5O Bruh I am so done with SF https://t.co/AJLyf2EB5Oview source ↗
-
• twitter 2d neutral$1B run rate was just the beginning $1B run rate was just the beginningview source ↗
-
• twitter 2d neutralHe’s 20, running a $50M AI memory startup backed by Google’s AI chief, Cloudflare’s CTO, and execs from OpenAI and Meta. A month ago he shut down half his products on purpose. This is Dhravya Shah, founder of Supermemory: the infrastructure that gives AI apps and agents long-term memory, so they actually remember you … He’s 20, running a $50M AI memory startup backed by Google’s AI chief, Cloudflare’s CTO, and execs from OpenAI and Meta. A month ago he shut down half his products on purpose. This is Dhravya Shah, founder of Supermemory: the infrastructure that gives AI apps and agents long-term memory, so they actually remember you between conversations. The company was making a few million a year, but the AI frontier would never use what he was selling and to him, a company that just makes money is a company that gets forgotten. So he went to his investors’ houses, told them he was killing a few million dollars in revenue, refunded every affected customer, and bet the company on building the memory that frontier AI will actually run on. The internship was two years ago. The $50M startup is now. If you're building, the episode is the playbook. Full episode releasing Wednesday on YouTube and Spotify.view source ↗
-
• twitter 2d positiveJust sell info bro ecom $100k is like $15k after meta, the product, shipping, refunds and chargebacks, and he has to do it all again tomorrow and pray the ad account doesnt get banned the same $100k selling info: - ads: $0, cuz content brings the people to you - product: $0, its what you already know - shipping: $0,… Just sell info bro ecom $100k is like $15k after meta, the product, shipping, refunds and chargebacks, and he has to do it all again tomorrow and pray the ad account doesnt get banned the same $100k selling info: - ads: $0, cuz content brings the people to you - product: $0, its what you already know - shipping: $0, its a zoom call - refunds: basically none when the guy is getting results and dont sell to cheapskate retards - stock: you never run out of yourself, its all online and in your head so you keep like 90%+ of it lmao and you dont even need a $100k day. 4 clients at $2-3k is a $10k month, and you can find 4 people with a few hundred followers. the ecom guy needs thousands of strangers clicking an ad every single month just to break even plus nobody can copy you. he finds a winning product and 50 chinese stores are selling it for half the price next week. nobody can sell YOUR experience cheaper than you, cuz people are paying for you and this isnt just for gurus btw. if you run a store, an agency or a 9-5 you already know shit people would pay to learn... you're just giving it away for free while you fight meta for a 15% margin -hhhview source ↗
-
• twitter 2d neutralStarting today, aura farming is accessible to everyone. With Higgsfield Katana: > Pick a preset or upload any edit > Add photos of your characters > Get your own version Connect the Higgsfield MCP and type /katana in Claude. https://t.co/X7Irj22kVb Starting today, aura farming is accessible to everyone. With Higgsfield Katana: > Pick a preset or upload any edit > Add photos of your characters > Get your own version Connect the Higgsfield MCP and type /katana in Claude. https://t.co/X7Irj22kVbview source ↗
-
• twitter 2d neutralwhole team feels dumb now, but it's true, we did not explicitly train our model on 3d demo games obviously in retrospect we probably should've done a bit of that... but no regrets, please try muse wonder on your actual coding work! whole team feels dumb now, but it's true, we did not explicitly train our model on 3d demo games obviously in retrospect we probably should've done a bit of that... but no regrets, please try muse wonder on your actual coding work!view source ↗
-
• twitter 2d negativeThe CIA trick behind his AI doomsday prediction. @NPCollapse puts the odds of AI taking control at 30% by 2027 - That’s a handful of months away - People get offended when he puts numbers on the end of the world - His answer is the rule CIA analysts are trained to follow “If you work for the CIA, you were trained… The CIA trick behind his AI doomsday prediction. @NPCollapse puts the odds of AI taking control at 30% by 2027 - That’s a handful of months away - People get offended when he puts numbers on the end of the world - His answer is the rule CIA analysts are trained to follow “If you work for the CIA, you were trained to not use words like likely or unlikely because this means totally different things to people. Instead, just put some number that feels right. And this is much more accurate.”view source ↗
-
• twitter 2d neutralCoaches Corner powered by @awscloud goes Inside the Game with San Antonio Spurs head coach Mitch Johnson and ESPN analyst Tim Legler as they explore how Victor Wembanyama's gravitational pull creates gaps and scoring opportunities for his teammates. https://t.co/dFU400M5aC Coaches Corner powered by @awscloud goes Inside the Game with San Antonio Spurs head coach Mitch Johnson and ESPN analyst Tim Legler as they explore how Victor Wembanyama's gravitational pull creates gaps and scoring opportunities for his teammates. https://t.co/dFU400M5aCview source ↗
-
• twitter 2d positiveTHIS IS ACTUAL F*CKING GOLD Anthropic just dropped one of the best tricks for Claude Code. The idea: run Sonnet, Haiku and Opus 5.5 together, each one doing only the job it's best at. Better output and way fewer tokens, because every task lands on the right model. You just launch Claude Code like this: claude --mo… THIS IS ACTUAL F*CKING GOLD Anthropic just dropped one of the best tricks for Claude Code. The idea: run Sonnet, Haiku and Opus 5.5 together, each one doing only the job it's best at. Better output and way fewer tokens, because every task lands on the right model. You just launch Claude Code like this: claude --model sonnet --advisor opus here's the split: → Sonnet 5 runs the session and writes the code → Haiku 5.5 explores files and docs in parallel → Opus 5.5 stays on call as advisor for the big calls → you stop burning Opus context on work that doesn't need it Opus only steps in at three moments: → before a plan locks → when the same error shows up twice → before calling it done Sonnet builds. Haiku explores. Opus reviews. JEV engineering is the same move one layer down: the forks that need no thinker (which file, which tool, retry or stop) go to JEV in milliseconds, and the big models only see the ones that split the full tree > Sonnet 5 runs the main session and edits code > search maps the repo on Haiku 5.5 > grep finds call sites on Haiku 5.5 > docs pulls the API reference on Haiku 5.5 > Opus 5.5 on call as advisor, reads the whole transcript when consulted > JEV routes micro-forks in milliseconds paste the tree and this prompt into Claude Code ↓ "Rebuild my Claude Code setup around this tree: 1. Run the main session on Sonnet 5 with Opus 5.5 as advisor. 2. Create three read-only Haiku 5.5 subagents: search (maps the repo), grep (finds call sites), docs (pulls the API reference). 3. Add one rule to CLAUDE.md: consult the advisor before locking a plan, after the same error twice, and before calling a task done. 4. Route fast deterministic forks (tool choice, path selection, retries) through JEV. Show me every change as a diff first. No edits until I say go." ↳ https://t.co/vOifNVjYxB you'll use this every single day.view source ↗
-
• twitter 2d positiveAs CEO of Product Hunt, I see hundreds of launches a week. One trait separates winners from losers. Every few months, new models come out. Some products become better, effortlessly. Some don't improve, or become irrelevant. This property needs a name. I'm calling it Buoyancy. Buoyant startups rise with model improve… As CEO of Product Hunt, I see hundreds of launches a week. One trait separates winners from losers. Every few months, new models come out. Some products become better, effortlessly. Some don't improve, or become irrelevant. This property needs a name. I'm calling it Buoyancy. Buoyant startups rise with model improvements the way a boat rises with the tide. Anti-buoyant startups come in two flavors: over-scaffolded (which hinders the new, smarter models) and too thin - which makes them redundant when models improve. To be buoyant, you need a "hull" - usually proprietary data, a vertical-specific harness, network effects, or distribution advantage. This hull keeps you buoyant as models improve. Hulls are the new moats. My number one advice to startups in 2026: Be buoyant. Find your hull. (Full essay on my blog)view source ↗
-
• twitter 2d neutraleveryone is calling muse benchmaxxed. it's uncalled for, used max thinking lvl for both. not very token friendly tho. are you getting similar results? https://t.co/5DMGmItlAP everyone is calling muse benchmaxxed. it's uncalled for, used max thinking lvl for both. not very token friendly tho. are you getting similar results? https://t.co/5DMGmItlAPview source ↗
-
• twitter 2d positiveit feels so good knowing i'll never buy another macbook pro in my life https://t.co/WoSOL4TUTb it feels so good knowing i'll never buy another macbook pro in my life https://t.co/WoSOL4TUTbview source ↗
-
• twitter 2d neutralSelma Blair was hospitalized and is in stable condition after allegedly injuring crew members aboard a Delta flight, police say https://t.co/dRdZbM4ttD https://t.co/FlBbLWHF6m Selma Blair was hospitalized and is in stable condition after allegedly injuring crew members aboard a Delta flight, police say https://t.co/dRdZbM4ttD https://t.co/FlBbLWHF6mview source ↗
-
• twitter 2d positiveOk @alexandr_wang you need to subsidize the subscription more. I signed up for the $50 plan to try it out, gave it a side project I was working on and ran out of subscription quota quickly. I am back using codex for that task. The model did look very good while it was working. Ok @alexandr_wang you need to subsidize the subscription more. I signed up for the $50 plan to try it out, gave it a side project I was working on and ran out of subscription quota quickly. I am back using codex for that task. The model did look very good while it was working.view source ↗
-
• twitter 2d neutral@RielStAmand Agreed. They’re fun but they don’t represent real work @RielStAmand Agreed. They’re fun but they don’t represent real workview source ↗
-
• twitter 2d positiveLOL, this app is making $12,000,000 in ARR It's an educational app And here's the viral UGC format behind their success The viral format → a student asks her AI tutor to explain a certain topic like juicy gossip, turning a boring lesson into something fun to listen to > The best video with this format got 1.7M view… LOL, this app is making $12,000,000 in ARR It's an educational app And here's the viral UGC format behind their success The viral format → a student asks her AI tutor to explain a certain topic like juicy gossip, turning a boring lesson into something fun to listen to > The best video with this format got 1.7M views Here’s the 5-step structure every top video follows: 1. Open with something students already relate to, like “studying like I’m listening to juicy gossip” 2. Let the tutor explain the topic the boring way first 3. Stop it and ask for the same lesson in a fun style 4. Let the fun version play 5. Show the StudyFetch app at the end Why it works: 1. The lesson is the fun part, so people learn what the app does while they laugh 2. Each video talks to one type of student, like anatomy students, which feels more personal than a video for everyone 3. Small creators reach more people than the brand. The official account gets a few thousand views at most, while creators often get tens of thousands 4. The comments tell them what to post next. When people ask “how do I do this?” or “is it free?”, that becomes the next video The format has 3 variations worth knowing: 1/ Make the lesson entertaining The tutor explains a hard topic as gossip, a story, or a funny roleplay. This one made the biggest hit 2/ Make studying a game Students play mini games inside the app, like nursing students checking on pretend patients or anatomy students quizzing themselves on a 3D heart 3/ Solve one class problem The video picks one hard class and highlights the one feature that helps. Students in that class see themselves right away and stop scrolling Key takeaways for marketers: 1. Make your app the reason the video is fun, so people learn what it does without feeling sold to 2. Make each video for one exact person, like “nursing students studying for adult health” 3. Show one feature per video, bcs showing everything at once looks like an ad 4. Turn comments into videoview source ↗
-
• twitter 2d negativeNever understood the "look at what this model did in Blender!" type of demo that people go nuts for. I have the same issue with that as I do "my agent books last minute travel!" demos. Doesn't represent real use cases in the slightest. Never understood the "look at what this model did in Blender!" type of demo that people go nuts for. I have the same issue with that as I do "my agent books last minute travel!" demos. Doesn't represent real use cases in the slightest.view source ↗
-
• twitter 2d positive@collision @finkd @alexandr_wang I was just pouring a pint there the other day! Best pub in silicon valley by far. https://t.co/Wjx7SUHFZe @collision @finkd @alexandr_wang I was just pouring a pint there the other day! Best pub in silicon valley by far. https://t.co/Wjx7SUHFZeview source ↗
-
• twitter 2d neutralYou won’t believe it until you experience it yourself. All from a one-line prompt. https://t.co/xLXqaZIEhu You won’t believe it until you experience it yourself. All from a one-line prompt. https://t.co/xLXqaZIEhuview source ↗
-
• twitter 2d positiveAlready see a lot of people discounting Muse Wonder because it wasn’t RL’d in 3d stuff like Astra and Opus. 3d game demos are cool! They don’t represent real code work at all. Wonder is an incredible code model. Don’t sleep on it. Already see a lot of people discounting Muse Wonder because it wasn’t RL’d in 3d stuff like Astra and Opus. 3d game demos are cool! They don’t represent real code work at all. Wonder is an incredible code model. Don’t sleep on it.view source ↗
-
• twitter 2d neutralsaid that last year, saying it again. vibes coding is a slur. said that last year, saying it again. vibes coding is a slur.view source ↗
-
• twitter 2d neutralMy current model rankings S - Opus 5.5, Muse Wonder A - GPT-6.1 Sol — Gap — Everything else My current model rankings S - Opus 5.5, Muse Wonder A - GPT-6.1 Sol — Gap — Everything elseview source ↗