•twitter5041
listening to
timeline
5041
mentions
tracked
tracked
net sentiment
+24%
1580 positive
3086 neutral
375 negative
volume & sentiment over time
Sep 30 – Oct 5, 2026 · Daily
sources
most discussed
—
active days
6
mentions
Showing 526–550 of 5041 mentions
-
• twitter 12h neutralRandom German lab appears from nowhere and drops SOTA open-weight model Beats Qwen, Mistral, Nematron models with similar active parameter count Trained from stratch https://t.co/2xY0qylFxa Random German lab appears from nowhere and drops SOTA open-weight model Beats Qwen, Mistral, Nematron models with similar active parameter count Trained from stratch https://t.co/2xY0qylFxaview source ↗
-
• twitter 12h positiveThere’s no point having a Usage Limit Reset every day if the models aren’t usable. I’m not saying Astra or 6.1 Sol are bad models, but once you use Opus 5.5, you don’t wanna use Astra or 6.1 Sol anymore… The best thing OpenAI can release and the only thing that would actually convince me to get a OpenAI sub again wo… There’s no point having a Usage Limit Reset every day if the models aren’t usable. I’m not saying Astra or 6.1 Sol are bad models, but once you use Opus 5.5, you don’t wanna use Astra or 6.1 Sol anymore… The best thing OpenAI can release and the only thing that would actually convince me to get a OpenAI sub again would be a new model that’s better than Opus 5.5 Good luck OpenAIview source ↗
-
• twitter 12h neutralAI IS SXSI NOW AI IS SXSI NOWview source ↗
-
• twitter 12h neutralI figured I'd touch some grass while CI builds... https://t.co/7WTJaxyWTj I figured I'd touch some grass while CI builds... https://t.co/7WTJaxyWTjview source ↗
-
• twitter 12h neutralExplore dangerous elemental temples in Kena: Scars of Kosmora New dev details highlight these challenging destinations and their transformative rewards: https://t.co/p7BdKzz0l0 https://t.co/5jBtUyMVY1 Explore dangerous elemental temples in Kena: Scars of Kosmora New dev details highlight these challenging destinations and their transformative rewards: https://t.co/p7BdKzz0l0 https://t.co/5jBtUyMVY1view source ↗
-
• twitter 12h positiveBeing loved is the minimum. Make sure you are also being respected, prioritized, supported, desired, and understood. Being loved is the minimum. Make sure you are also being respected, prioritized, supported, desired, and understood.view source ↗
-
• twitter 12h neutralFormer Anthropic researcher Coxon to testify at New York City AI hearing, Bloomberg News reports https://t.co/Vrcgcb7SBq https://t.co/Vrcgcb7SBq Former Anthropic researcher Coxon to testify at New York City AI hearing, Bloomberg News reports https://t.co/Vrcgcb7SBq https://t.co/Vrcgcb7SBqview source ↗
-
• twitter 12h positiveA man found a small piece of paper under his door. It said: “Don’t give up today.” No name. No explanation. He laughed, threw it away, and went about his day. That evening, he found out the person who had written the note had passed away that morning. The strange part? They had never met. Sometimes, a message r… A man found a small piece of paper under his door. It said: “Don’t give up today.” No name. No explanation. He laughed, threw it away, and went about his day. That evening, he found out the person who had written the note had passed away that morning. The strange part? They had never met. Sometimes, a message reaches you at exactly the moment you need it. ❤️view source ↗
-
• twitter 12h positiveI just ported 'npx RuFlo' to the new Claude Code Mod system. It's next level. I was inspired to recreate a 1990 Wildcat BBS I ran back in the day. A mod is a Claude Code plugin whose behaviour is a hooks module: JavaScript or TypeScript that exports register(on, options). Inside it, on(event, hook) attaches middlewar… I just ported 'npx RuFlo' to the new Claude Code Mod system. It's next level. I was inspired to recreate a 1990 Wildcat BBS I ran back in the day. A mod is a Claude Code plugin whose behaviour is a hooks module: JavaScript or TypeScript that exports register(on, options). Inside it, on(event, hook) attaches middleware to the engines own events. Each hook can watch an event, rewrite it, or answer it in place of the engine. @claudeai can also directly control the console and the console can directly steer claude code autonomously. Steps to install and use it in claude code cli. /plugin marketplace add ruvnet/ruflo /plugin install ruflo-console@ruflo /plugin install ruflo-mods@ruflo /reload-plugins /ruflo ** make sure you're using the latest claude cli or desktop. If you like it, ⭐ it. https://t.co/lptBKeOT3Y cc @bcherny @jackclarkSF @ClaudeDevsview source ↗
-
• twitter 12h neutralnever bet against a guy who beat cancer at 24, ditched the M&A lawyer track and went all-in in SF to build Strava for work never bet against a guy who beat cancer at 24, ditched the M&A lawyer track and went all-in in SF to build Strava for workview source ↗
-
• twitter 12h positiveThe best developers I know aren’t just fast typists. They are ruthless about what not to build. The best developers I know aren’t just fast typists. They are ruthless about what not to build.view source ↗
-
• twitter 12h positiveSetting up a nice home office now, what have u guys found that has made a huge difference? I want to upgrade my standing desk, buy new monitor arms etc but ideally something nice for recording videos etc Setting up a nice home office now, what have u guys found that has made a huge difference? I want to upgrade my standing desk, buy new monitor arms etc but ideally something nice for recording videos etcview source ↗
-
• twitter 12h neutralMessi touches down in Argentina as a member of the national team for the last time ever 🇦🇷🐐 (via @Argentina) https://t.co/wQfK0TnC7t Messi touches down in Argentina as a member of the national team for the last time ever 🇦🇷🐐 (via @Argentina) https://t.co/wQfK0TnC7tview source ↗
-
• twitter 12h positiveI'm building with @rive_app , a lightweight app to render and interact with any agent workflow in real time, abstracting away the conversation session layer. It's opinionated on the edge but deeply customizable inside. To iterate on the designs, I’m deploying a fleet of Claude Sonnet 5.5 workers orchestrated by Claud… I'm building with @rive_app , a lightweight app to render and interact with any agent workflow in real time, abstracting away the conversation session layer. It's opinionated on the edge but deeply customizable inside. To iterate on the designs, I’m deploying a fleet of Claude Sonnet 5.5 workers orchestrated by Claude Opus 5.5. They generate variations of an RML template constrained by a custom design system and a layout concept to stay aligned with the overall desired aesthetic. Here is what some of these variations look like using the 'strip' layout concept:view source ↗
-
• twitter 12h neutralsup nerds we're live https://t.co/C400PYa1Fv sup nerds we're live https://t.co/C400PYa1Fvview source ↗
-
• twitter 12h negativeI think I have a problem... https://t.co/C400PYa1Fv I think I have a problem... https://t.co/C400PYa1Fvview source ↗
-
• twitter 12h positive@ryeggz @alexhvnsen @dhh Here's link for headless server. We're absolutely aware that's too much to handle. ETA is next Wednesday, for a simple "it just works" setup and updates! Ping you guys here when it comes out, if that's ok?? 🫶 https://t.co/mrTcFyqnjw @ryeggz @alexhvnsen @dhh Here's link for headless server. We're absolutely aware that's too much to handle. ETA is next Wednesday, for a simple "it just works" setup and updates! Ping you guys here when it comes out, if that's ok?? 🫶 https://t.co/mrTcFyqnjwview source ↗
-
• twitter 12h neutralhttps://t.co/FCxmrKSlW8 https://t.co/FCxmrKSlW8view source ↗
-
• twitter 12h positive@ryeggz @alexhvnsen @dhh Thanks for the shout out, Alex, and for the opportunity of trying us out, Ryan ❤️ In headed mode, Orca Apps can be accessed from other Orca as clients, desktop and mobile! Although still a bit verbose, headless server is there too. (1/2) https://t.co/hMQUrZOVMr @ryeggz @alexhvnsen @dhh Thanks for the shout out, Alex, and for the opportunity of trying us out, Ryan ❤️ In headed mode, Orca Apps can be accessed from other Orca as clients, desktop and mobile! Although still a bit verbose, headless server is there too. (1/2) https://t.co/hMQUrZOVMrview source ↗
-
• twitter 12h neutralThis Former Disney Princess Lived ‘Paycheck to Paycheck’ Before Starting a Side Hustle at Home — Now She Makes $250,000 a Year Read more: https://t.co/Hz9JZKnX5d This Former Disney Princess Lived ‘Paycheck to Paycheck’ Before Starting a Side Hustle at Home — Now She Makes $250,000 a Year Read more: https://t.co/Hz9JZKnX5dview source ↗
-
• twitter 12h neutralhttps://t.co/yOp8d1R5CT https://t.co/yOp8d1R5CTview source ↗
-
• twitter 12h positiveThe key question: Will Anthropic/OpenAI gain a durable advantage via recursive self-improvement (RSI)? Or will AI models become a commodity? I think the top labs will continue to stay a few months ahead of everyone else. But two things limit RSI: - Compute can't scale quickly in the near term (hardware singularity … The key question: Will Anthropic/OpenAI gain a durable advantage via recursive self-improvement (RSI)? Or will AI models become a commodity? I think the top labs will continue to stay a few months ahead of everyone else. But two things limit RSI: - Compute can't scale quickly in the near term (hardware singularity still years off) - Intelligence saturation: Marginal returns to more intelligence will start *declining* (even though they previously were *increasing*) for most problems The top labs have an upper bound on their progress from compute limits, which prevents a true takeoff. At the same time, AI agents are saturating benchmarks for chat, research, computer use, spreadsheet work, and soon, I hypothesize, coding. If this is true, then new flagship models will be better, but not enough better to command much higher prices than now. Meanwhile, the diffusion of AI research will continue in full force, allowing all the other labs to stay close in capabilities. So, the price of intelligence will continue to drop incredibly, and cheap models will be able to do most of the work people want done. Most people will not pay a subscription for that but get it free, funded by advertising, on the same business model that allowed Google to offer so many Web 2.0 products for free. ------- For the sign of the marginal return to more intelligence, a key question is how more intelligence affects AI research itself. Getting to the point where AI can execute valuable experiments reliably to fill compute-time is incredibly valuable. But, we're pretty much there today. After all the compute is being used for reasonable experiments with a certain expected gain each, how much could more intelligence help? We'd need AI to help with research taste. Plausibly, new top models *could* make important contributions here, which gives some uncertainty to the overall thesis. But knowing that every other benchmark or task follows the same saturation curve means that even AI researview source ↗
-
• twitter 12h neutralKeep believing. Work hard. Your time will come. Keep believing. Work hard. Your time will come.view source ↗
-
• twitter 12h neutralPeople will judge you no matter what. So why even bother what they think or say? https://t.co/vhZzLFGPOD People will judge you no matter what. So why even bother what they think or say? https://t.co/vhZzLFGPODview source ↗
-
• twitter 12h neutralhttps://t.co/z4gizp4kc8 https://t.co/z4gizp4kc8view source ↗