•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 2926–2950 of 4998 mentions
-
• twitter 1w positiveIn 1984, a researcher found that students with a 1:1 tutor did better than 98% of a regular class. we've known the answer for 40 years. we just couldn't give everyone a tutor. for language learners, now we can. Step-by-step guide: https://t.co/2ZUtcfg1XN In 1984, a researcher found that students with a 1:1 tutor did better than 98% of a regular class. we've known the answer for 40 years. we just couldn't give everyone a tutor. for language learners, now we can. Step-by-step guide: https://t.co/2ZUtcfg1XNview source ↗
-
• twitter 1w positiveAnthropic is giving away $250 in free Claude Code credits right now. 2 days left to claim A one-time credit for cloud sessions. > Max gets $250, Pro gets $100 > Existing subscribers only > Claim by Oct 7, 11:59 PM PT > Expires Nov 4, no matter when you claim > Needs a GitHub repo connected > Works in cloud sess… Anthropic is giving away $250 in free Claude Code credits right now. 2 days left to claim A one-time credit for cloud sessions. > Max gets $250, Pro gets $100 > Existing subscribers only > Claim by Oct 7, 11:59 PM PT > Expires Nov 4, no matter when you claim > Needs a GitHub repo connected > Works in cloud sessions only Claim link: https://t.co/WFpsdbc4iL Or open the Usage tab in your Claude settingsview source ↗
-
• twitter 1w positive@nizzyabi most of the "hype capital" gets deployed to the layer closest to the end user, making it also the most volatile a layer down, in infra, its much more sensible also competitive but you can specialise on a set of use cases and be the best in one thing and let the market help you @nizzyabi most of the "hype capital" gets deployed to the layer closest to the end user, making it also the most volatile a layer down, in infra, its much more sensible also competitive but you can specialise on a set of use cases and be the best in one thing and let the market help youview source ↗
-
• twitter 1w positiveRT @attacomsian: thank you @ProductHunt for featuring Vitals in your weekly newsletter 🙏 I built it in two days to fix my own Mac. now 1,2… RT @attacomsian: thank you @ProductHunt for featuring Vitals in your weekly newsletter 🙏 I built it in two days to fix my own Mac. now 1,2…view source ↗
-
• twitter 1w positiveI've been playing around with this speech to text model. It's incredibly capable. Amazing that it all fits in 17MB. That's an entire speech to text neural network squeezed down to the size of an MP3 of Hey Jude. I've been playing around with this speech to text model. It's incredibly capable. Amazing that it all fits in 17MB. That's an entire speech to text neural network squeezed down to the size of an MP3 of Hey Jude.view source ↗
-
• twitter 1w positiveRT @EXM7777: i spent 70+ hours with Opus 5.5 inside Claude Code... here's how to get the best out of it: - set effort to medium at all tim… RT @EXM7777: i spent 70+ hours with Opus 5.5 inside Claude Code... here's how to get the best out of it: - set effort to medium at all tim…view source ↗
-
• twitter 1w positive$147 TODAY ON LAUNCHIT 🚀 Three sales. All from the same customer. He ordered three blog reviews at once, one for each of his websites. When someone trusts you enough to buy three on the first go, you know the offer makes sense. https://t.co/vB4MzOFR51 $147 TODAY ON LAUNCHIT 🚀 Three sales. All from the same customer. He ordered three blog reviews at once, one for each of his websites. When someone trusts you enough to buy three on the first go, you know the offer makes sense. https://t.co/vB4MzOFR51view source ↗
-
• twitter 1w positivedon't WASTE your time... this is a once-in-a-lifetime opportunity there are literally campaigns paying you to create viral AI influencers these accounts are going viral everywhere, getting 77M+ views in just a few days now you can make $1 per 1,000 views from them, here's how: - create your own character with Higgsf… don't WASTE your time... this is a once-in-a-lifetime opportunity there are literally campaigns paying you to create viral AI influencers these accounts are going viral everywhere, getting 77M+ views in just a few days now you can make $1 per 1,000 views from them, here's how: - create your own character with Higgsfield AI Influencer - study viral formats from creators like Jean Pill, Derek Mencer, etc. - use Genjutsu to adapt those formats to your character - upload your videos and watch the views roll in congrats, now you have a video with 100M+ view potentialview source ↗
-
• twitter 1w positiveIntroducing INTI: a cloud computer for coding agents that ship mobile apps for free - Sandbox w/ 4 CPU · 8 GB RAM · 10 GB disk - Runs opencode, claude code, pi or codex - SSH in or w/ Ghostty web terminal (both stay in sync) - iOS cloud sims CIO: https://t.co/8o6KYFnyth https://t.co/TbSnGc4LLA Introducing INTI: a cloud computer for coding agents that ship mobile apps for free - Sandbox w/ 4 CPU · 8 GB RAM · 10 GB disk - Runs opencode, claude code, pi or codex - SSH in or w/ Ghostty web terminal (both stay in sync) - iOS cloud sims CIO: https://t.co/8o6KYFnyth https://t.co/TbSnGc4LLAview source ↗
-
• twitter 1w positiveThis is actually super helpful! There are multiple things this helped clarify that wasn't obvious to me: it is the Ghostty app (not libghostty), it is directly via `cat`, it is on macOS. Those are all really critical details and that helps, thank you. Is your app available (not sure what app that is) so I can reprodu… This is actually super helpful! There are multiple things this helped clarify that wasn't obvious to me: it is the Ghostty app (not libghostty), it is directly via `cat`, it is on macOS. Those are all really critical details and that helps, thank you. Is your app available (not sure what app that is) so I can reproduce these results? No need for source of course, just so I can see the relative numbers with the same inputs on the array of Apple devices I own.view source ↗
-
• twitter 1w positiveDr Karpathy tweeted about this yesterday... First time I heard about "ASD-STE100". Asked Claude to explain what it is in an HTML... Its really cool that this style of writing has a name! https://t.co/v3dIKe0CAW Dr Karpathy tweeted about this yesterday... First time I heard about "ASD-STE100". Asked Claude to explain what it is in an HTML... Its really cool that this style of writing has a name! https://t.co/v3dIKe0CAWview source ↗
-
• twitter 1w positiveso turns out better web search results in better rl outcomes who would have thought? i personally think frontier models searching the internet themselves is kinda dumb a lot of the time if any researchers need search/scraping credits hmu. we’re about to see a lot of small, specialized searchers so turns out better web search results in better rl outcomes who would have thought? i personally think frontier models searching the internet themselves is kinda dumb a lot of the time if any researchers need search/scraping credits hmu. we’re about to see a lot of small, specialized searchersview source ↗
-
• twitter 1w positiveLet's goo 🚀🚀 GLM 5.3 Flash absolutely rocks now with TensorFold! Coming soon: TP=4 path (4x DGX Sparks) Get it here: https://t.co/fQzcacQtu5 Let's goo 🚀🚀 GLM 5.3 Flash absolutely rocks now with TensorFold! Coming soon: TP=4 path (4x DGX Sparks) Get it here: https://t.co/fQzcacQtu5view source ↗
-
• twitter 1w positiveThe theory of reflexivity and how it's causing markups “You have a company that does well, the revenue is growing really fast, and it has a quick markup. Because of that quick markup, it gets more capital, more notoriety, more human capital, and grows faster. Then a copycat investor might say, ‘If it is marked up, it… The theory of reflexivity and how it's causing markups “You have a company that does well, the revenue is growing really fast, and it has a quick markup. Because of that quick markup, it gets more capital, more notoriety, more human capital, and grows faster. Then a copycat investor might say, ‘If it is marked up, it is going to be marked up again.’ Then reflexivity kicks in, and everybody starts acting that way. All reflexivity will eventually stop. We just do not know how and when.” @venkyganesan You see this every single day with your companies @rabois @chetanp @shaunmaguire how do you advise founders on when to take the money on the table vs when capital itself is not the weapon/moat?view source ↗
-
• twitter 1w positiveMy whole feed is AI agents right now. Last week, 100k people found out I'd never used the one everyone calls the best. A guy ranked Hermes Agent S tier. Grok bot and Muse, the two agents I actually live in, got A. I quoted it and asked what I was missing. It blew up. So I stopped asking and went and looked properl… My whole feed is AI agents right now. Last week, 100k people found out I'd never used the one everyone calls the best. A guy ranked Hermes Agent S tier. Grok bot and Muse, the two agents I actually live in, got A. I quoted it and asked what I was missing. It blew up. So I stopped asking and went and looked properly. The first thing I noticed: the loudest complaint about Hermes, the "$1,000 a month" bills, mostly isn't about Hermes. It's about people running their whole life through one chat. Every message drags the full thread along with it, so you pay for every old conversation again and again. Hermes Desktop breaks that into sessions you can pin and file into folders. One topic, one session. Boring fix. Big difference. Then I hit profiles, and this is where I disagree with a lot of the hype. Each profile is its own agent with its own skills, memory and personality. The trend is to build a fake company out of them. CEO agent. PM agent. Designer agent. Intern agent. It looks great in a screenshot. In practice, every handoff re-reads context, and you end up paying four agents to make one thumbnail. The setup that makes sense to me is one profile per model. The smartest, most expensive one plans. GPT codes, because the limits are bigger. A local model does research, because it costs nothing. I'd rather pick the model for the job than the job title. A few things caught me off guard. Hermes ships with 150+ skills, and every one quietly sits in your context and runs up the bill. There's now a panel to switch off the ones you never use. It also writes new skills from the things you repeat, so opening that tab feels a bit like reading your agent's private notes about you. Scheduled tasks are finally something I'd trust. In chat apps I'd type "do this every morning at 9" and just hope. Now there's a tab listing every job that actually exists. The best trick I picked up is to dump everything you care about into the chat, then ask it to write the best prompt forview source ↗
-
• twitter 1w positivePi Durable is awesome! Asked Opus 5.5 to make a deep technical manual on it. It's really good and the diagrams work really well to explain things 🤌 https://t.co/SBjAR5Zcxj https://t.co/R9KKka2pUI Pi Durable is awesome! Asked Opus 5.5 to make a deep technical manual on it. It's really good and the diagrams work really well to explain things 🤌 https://t.co/SBjAR5Zcxj https://t.co/R9KKka2pUIview source ↗
-
• twitter 1w positiveGood lighting and a good mic do more for your video calls than a new laptop. My home office list has the key light, podcast mic, 4K monitor and Wi-Fi 7 mesh I'd start with, and some are already marked down. #PrimeBigDealDays @amazon (affiliate link) https://t.co/BjYUx5i0QF Good lighting and a good mic do more for your video calls than a new laptop. My home office list has the key light, podcast mic, 4K monitor and Wi-Fi 7 mesh I'd start with, and some are already marked down. #PrimeBigDealDays @amazon (affiliate link) https://t.co/BjYUx5i0QFview source ↗
-
• twitter 1w positiveMessi leading out the national team will forever be iconic 🔥🇦🇷 https://t.co/KnFRqnXJCx Messi leading out the national team will forever be iconic 🔥🇦🇷 https://t.co/KnFRqnXJCxview source ↗
-
• twitter 1w positiveholy sauce alert: seasonal creatives been testing these i had opus make and i am seeing 4x ROAS on these, never seen anything quite like it. https://t.co/TGr1eMdpYv holy sauce alert: seasonal creatives been testing these i had opus make and i am seeing 4x ROAS on these, never seen anything quite like it. https://t.co/TGr1eMdpYvview source ↗
-
• twitter 1w positiveIt finally happened America has entered the frontier open weights model competition. We aren't rolling over to China Reflection announced Beam, a 501B parameter open weights model that is comparable to GLM 5.2, Qwen 3.8, and Opus 4.8 Yes, Opus 4.8 on your desk You will need a decent sized machine for this. Probably… It finally happened America has entered the frontier open weights model competition. We aren't rolling over to China Reflection announced Beam, a 501B parameter open weights model that is comparable to GLM 5.2, Qwen 3.8, and Opus 4.8 Yes, Opus 4.8 on your desk You will need a decent sized machine for this. Probably the Mac Studio 512gb But if you listened to me back in January when I warned you days like today were going to come, then you're all set I stand by Opus 4.5 being the most important model release in history. We will now have a model BETTER than that on our desks If you haven't gotten into local AI yet, it's officially time to do it Take any computer you have, talk to an AI agent, ask which models you can run Never been more important to just startview source ↗
-
• twitter 1w positiveif you want self-improving software, you need to fit your app in a context graph that's the only way to enable agents to understand everything you run I personally built this over the past 6 months, here are all the learnings, feel free to copy https://t.co/uFkhstkASg if you want self-improving software, you need to fit your app in a context graph that's the only way to enable agents to understand everything you run I personally built this over the past 6 months, here are all the learnings, feel free to copy https://t.co/uFkhstkASgview source ↗
-
• twitter 1w positiveDay 51 of scaling my app to $5k MRR. $1,000 MRR crossed. Exactly 2 months since Day 1. $182 → $1,010 MRR. The biggest difference has been treating distribution like a system. Test formats → track what actually drives installs & revenue → find the patterns → replicate what works. Not just chasing views. That proc… Day 51 of scaling my app to $5k MRR. $1,000 MRR crossed. Exactly 2 months since Day 1. $182 → $1,010 MRR. The biggest difference has been treating distribution like a system. Test formats → track what actually drives installs & revenue → find the patterns → replicate what works. Not just chasing views. That process is starting to compound now, while I’m putting most of my time into improving the app and conversion. Still a long way to $5k, but the direction is clear. Time to double down.view source ↗
-
• twitter 1w positiveI particularly like the call stacks (got this from @dillon_mulroy) and ability to annotate code snippets https://t.co/t5s7hSdSdx I particularly like the call stacks (got this from @dillon_mulroy) and ability to annotate code snippets https://t.co/t5s7hSdSdxview source ↗
-
• twitter 1w positiveExcited about this loop! Your docs should be useful to you, your team, AND your agents - without forcing everyone into a new file system or collaboration model. :) Excited about this loop! Your docs should be useful to you, your team, AND your agents - without forcing everyone into a new file system or collaboration model. :)view source ↗
-
• twitter 1w positiveGenjutsu + character = crazy content loop 🔥 Genjutsu + character = crazy content loop 🔥view source ↗