•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 3326–3350 of 4998 mentions
-
• twitter 1w positiveCool, calm, and collected 🥶 Late drama as Milner scored from the spot to secure all three points against the Foxes ⏪ 05/10/19 🗓️ https://t.co/RyGP4BleZ6 Cool, calm, and collected 🥶 Late drama as Milner scored from the spot to secure all three points against the Foxes ⏪ 05/10/19 🗓️ https://t.co/RyGP4BleZ6view source ↗
-
• twitter 1w positiveStrength is the ability to remain useful when the room becomes difficult. Strength is the ability to remain useful when the room becomes difficult.view source ↗
-
• twitter 1w positive🚨 Codex Users 🚨 GPT-6.1 Sol can create some amazing motion designs. I asked Dots to create a presentation about itself using GPT-6.1 Sol with xHigh effort. It took 43 minutes, and this was the result: 👇 https://t.co/x4O569VtIe 🚨 Codex Users 🚨 GPT-6.1 Sol can create some amazing motion designs. I asked Dots to create a presentation about itself using GPT-6.1 Sol with xHigh effort. It took 43 minutes, and this was the result: 👇 https://t.co/x4O569VtIeview source ↗
-
• twitter 1w positiveStruggling to sit down to code on my laptop… This mobile app is sooo fkn good https://t.co/sPkMQ62NUu Struggling to sit down to code on my laptop… This mobile app is sooo fkn good https://t.co/sPkMQ62NUuview source ↗
-
• twitter 1w positiveLots of praise for Google’s Gemini 4 Argon. But is it actually better at coding than Opus or Astra? Or did we just get really good at benchmarking? Lots of praise for Google’s Gemini 4 Argon. But is it actually better at coding than Opus or Astra? Or did we just get really good at benchmarking?view source ↗
-
• twitter 1w positiveYou can't lose bro. You are the last hope of your parents. You have to win. You can't lose bro. You are the last hope of your parents. You have to win.view source ↗
-
• twitter 1w positiveI’ve been switching between Claude models more and more. Claude writes the code. Claude reviews the code. Claude fixes the code. At this point, I’m just sitting there approving PRs, I didn’t write. Either Claude got insanely good, or I’ve completely forgotten how to code. Probably both. I’ve been switching between Claude models more and more. Claude writes the code. Claude reviews the code. Claude fixes the code. At this point, I’m just sitting there approving PRs, I didn’t write. Either Claude got insanely good, or I’ve completely forgotten how to code. Probably both.view source ↗
-
• twitter 1w positiveDevelopers, Which OS would you recommend every developer to experience once? - Linux - Windows - macOS https://t.co/OmHEFx3n3R Developers, Which OS would you recommend every developer to experience once? - Linux - Windows - macOS https://t.co/OmHEFx3n3Rview source ↗
-
• twitter 1w positivePremium Clean app screenshot design https://t.co/kja5Hnd7bZ Premium Clean app screenshot design https://t.co/kja5Hnd7bZview source ↗
-
• twitter 1w positive4:30AM wake up call! Great way to start the week. Let’s get it! https://t.co/kqbblTM1p3 4:30AM wake up call! Great way to start the week. Let’s get it! https://t.co/kqbblTM1p3view source ↗
-
• twitter 1w positiveGreat explainer on e2e, a testing framework by @TesterArmy powered by @agent_device on mobile Great explainer on e2e, a testing framework by @TesterArmy powered by @agent_device on mobileview source ↗
-
• twitter 1w positiveSpace Bunny turns an 8-hour workday into a 1-hour agent wall from one prompt. it splits the workload into 5 parallel branches: inbox, reports, data, calls and planning, then runs them at the same time inside one self-contained HTML file. 47 emails clear in one pass, the weekly report finishes in 90 seconds, and the d… Space Bunny turns an 8-hour workday into a 1-hour agent wall from one prompt. it splits the workload into 5 parallel branches: inbox, reports, data, calls and planning, then runs them at the same time inside one self-contained HTML file. 47 emails clear in one pass, the weekly report finishes in 90 seconds, and the data branch moves through 840 files an hour. instead of chaining tasks one by one, the model keeps all 5 branches moving while the dashboard tracks exactly where the time goes. 8 hours → 1 hour. 5 branches. 1 prompt. this is what agent workflows look like when parallelism is the default.view source ↗
-
• twitter 1w positivethis is like the biggest lie ever told and might end arena lol https://t.co/jvan68rep0 this is like the biggest lie ever told and might end arena lol https://t.co/jvan68rep0view source ↗
-
• twitter 1w positiveThis AGENTS.md fixes the three most annoying habits of Codex and Claude Code: - Writing backward-compatible code for a project nobody uses yet - Keeping old code and docs it no longer needs - Adding a pile of tests just to change a button color After months of daily use, these are the rules I wish I'd added from day … This AGENTS.md fixes the three most annoying habits of Codex and Claude Code: - Writing backward-compatible code for a project nobody uses yet - Keeping old code and docs it no longer needs - Adding a pile of tests just to change a button color After months of daily use, these are the rules I wish I'd added from day one: if the project has no real users yet, skip compatibility and migrations. Build the simplest code that works. Then follow four steps: docs first → tests → smallest implementation → clean up. Copy and paste it into your AGENTS.md 👇view source ↗
-
• twitter 1w positiveBest setup I've had to date 🤩 https://t.co/rk5GOIqzpe Best setup I've had to date 🤩 https://t.co/rk5GOIqzpeview source ↗
-
• twitter 1w positiveStay hopeful. You will laugh again, feel happy again and be glad you kept going. Stay hopeful. You will laugh again, feel happy again and be glad you kept going.view source ↗
-
• twitter 1w positivethe @poteto vs @thsottiaux reply section as a fighting game likes = damage, who's gonna win? my bot checks every hours so video continues https://t.co/a09Cvrej42 the @poteto vs @thsottiaux reply section as a fighting game likes = damage, who's gonna win? my bot checks every hours so video continues https://t.co/a09Cvrej42view source ↗
-
• twitter 1w positiveWhile building your B2B outbound strategy from scratch, turn the entire market into a set of hypotheses. Before you build a list, answer these: 1/ Who is actually worth targeting? Define the situation they’re in. - recently funded. - hiring their first sales team. - expanding into a new market. - replacing a compet… While building your B2B outbound strategy from scratch, turn the entire market into a set of hypotheses. Before you build a list, answer these: 1/ Who is actually worth targeting? Define the situation they’re in. - recently funded. - hiring their first sales team. - expanding into a new market. - replacing a competitor. - growing without adding headcount. this way, you’re defining who is most likely to have the problem. 2/ Why should they care right now? A company can be a perfect fit and still have 0 reason to talk to you today. Look for something that changed. - new hire - funding round - team expansion - new product - leadership change - shift in their tech stack The goal is to find the reason this problem matters now, not 6 months from now. 3/ Who actually feels the problem? Ask: - who deals with this problem every week? - who owns the KPI it affects? - who gets blamed when it goes wrong? - who has enough influence to fix it? Sometimes that’s the VP, sometimes it’s the person 2 levels below them. 4/ What should make them reply? Before writing the message, write the reply you want. “yeah, we’re dealing with this.” “we’re actually looking at this right now.” “how are you doing that?” “send me more.” Then work backwards. What problem, trigger or observation would make that reply feel natural? If you can’t answer that, your copy probably isn’t the problem yet. 5/ How should each segment be approached differently? If every segment gets the same message, you haven’t really segmented anything. A newly funded company should get a different angle from a company cutting headcount, vs a new VP should get a different message from someone who’s been in the role for five years. The signal should change the campaign otherwise it’s just personalization theatre. 6/ What did the campaign teach you? Track the hypothesis behind the campaign. Which situations produced the best conversations? Which signals actually predicted interest? Which personas moved fview source ↗
-
• twitter 1w positivecurrently @t3dotcodes is the best AI coding app for multitasking because of 2 main reasons: 1. the chats in the sidebar have actual useful info - Codex and Claude Code's default sidebar is focused on saving space and Codex's activity bar hides chats that i haven't archived yet so it's not usable by itself 2. threads … currently @t3dotcodes is the best AI coding app for multitasking because of 2 main reasons: 1. the chats in the sidebar have actual useful info - Codex and Claude Code's default sidebar is focused on saving space and Codex's activity bar hides chats that i haven't archived yet so it's not usable by itself 2. threads that are in progress are hidden - this saves so much mental load scanning the sidebar to see what's going on please copy the T3 Code sidebar 🙏 the Codex, Claude Code, and Cursor sidebars only show the title and a small coloured dot for some reason - it feels like they're all focused on design and being minimalistic/saving space this small coloured status dot trend needs to go badly @DevinAI is another example of a good sidebar at the moment, it shows stuff like "Approve phase 1" or "PR is ready" so you can EASILY scan the sidebar and see what's happeningview source ↗
-
• twitter 1w positive“You need to go to the gym to get jacked” is a lie. Looking lean is mostly diet. We already have muscles, just hidden by layers of fat. Of course you need to grow muscles to look jacked, but you don’t need to bench 100kg. Your body weight is already weight. Calisthenics will make you jacked, at home, for free, witho… “You need to go to the gym to get jacked” is a lie. Looking lean is mostly diet. We already have muscles, just hidden by layers of fat. Of course you need to grow muscles to look jacked, but you don’t need to bench 100kg. Your body weight is already weight. Calisthenics will make you jacked, at home, for free, without going to the gym with a PT. And you need some form of cardio. Walking is great. Running, cycling, swimming are even better. It doesn’t have to be intense and, like calisthenics, it’s mostly free. My fitness journey started with calisthenics at home, and surfing for cardio. I upgraded with 15kg dumbbells, and only started going to the gym last year. Start with what’s convenient. Make it a habit. Upgrade from there.view source ↗
-
• twitter 1w positiveGet late invoices paid without chasing clients yourself. 59% of small businesses have invoices more than 30 days overdue, and the average business is owed $17.7K. My bet: most of that money just never got asked for twice, because asking feels awkward. Claude doesn't feel awkward. The setup is four steps and about 20 … Get late invoices paid without chasing clients yourself. 59% of small businesses have invoices more than 30 days overdue, and the average business is owed $17.7K. My bet: most of that money just never got asked for twice, because asking feels awkward. Claude doesn't feel awkward. The setup is four steps and about 20 minutes. First, connect. Go to Settings, Connectors, Browse, and connect the tool your invoices live in (QuickBooks, Xero or Stripe) plus Gmail. Claude now sees who owes you what, live, instead of you exporting a spreadsheet. Then teach it your tone. Paste three reminder emails you've actually sent and type "turn this into a skill called /chase-invoices". Add your payment terms, your late fee and one rule: it gets firmer with time. Day 3 gets a friendly nudge with the pay link. By day 14 it's a firm reminder with the invoice re-attached, and at day 30 a final notice quotes your late fee. At day 45 it stops writing and tells you to pick up the phone. Then schedule it. Click Scheduled, New task, Weekly, Monday at 8:00, and give it one line: "Run /chase-invoices on every overdue invoice, drafts only." It runs on Claude's side, so it works with your laptop closed. Last, approve. Monday morning you open one list: who's late, by how many days, and a draft for each client in your voice. Nothing goes out until you approve it. Read, tweak, send. Five minutes. Two rules I'd add before the first Monday. Tell it to check Stripe for payments made since the invoice went out, or you'll chase a client who paid yesterday. And keep a do-not-chase list for anyone on a payment plan or in a dispute. Set it up this week, before you write another "just following up" email.view source ↗
-
• twitter 1w positivejust compared 18 places to build a startup. France ranks 17th 😬 with €100K profit, you get: Dubai: €99,187 Singapore: €91,796 Zug: €85,514 Lisbon: €71,488 San Francisco: €67,183 Paris: €58,757 in France the debate is always the same: "with this tax, will the millionaires leave?" I think it's the wrong debate. the … just compared 18 places to build a startup. France ranks 17th 😬 with €100K profit, you get: Dubai: €99,187 Singapore: €91,796 Zug: €85,514 Lisbon: €71,488 San Francisco: €67,183 Paris: €58,757 in France the debate is always the same: "with this tax, will the millionaires leave?" I think it's the wrong debate. the millionaires will be fine. the real question is where the next generation of talented 25-year-olds will build their products. they go everywhere they want no strings attached I'm French and I'm staying but I want France to be the place founders move to made a free tool so you can run your own numbers 👉 should i be https://t.co/uQDZjKsFh7 ???view source ↗
-
• twitter 1w positiveThis morning with OpenAI: 30% - "The AI service is temporarily overloaded. Please try again in a moment" 20% - "I can't honestly help you with [...task]" 40% - Super stupid and does sloppy work. 10% - Actual good work done. While limits leaking like crazy. Nerfed and rugpulled. This morning with OpenAI: 30% - "The AI service is temporarily overloaded. Please try again in a moment" 20% - "I can't honestly help you with [...task]" 40% - Super stupid and does sloppy work. 10% - Actual good work done. While limits leaking like crazy. Nerfed and rugpulled.view source ↗
-
• twitter 1w positiveI came from Hermes. It's free to download. Grok Bot needs a paid plan. I'd still pay for Grok Bot every time. Here's the honest scorecard 👇 GROK BOT WINS No server to run. Download the app and go Your Bots work on a cloud computer with its own browser It hands YOU the controls for passwords, 2FA and payments You set … I came from Hermes. It's free to download. Grok Bot needs a paid plan. I'd still pay for Grok Bot every time. Here's the honest scorecard 👇 GROK BOT WINS No server to run. Download the app and go Your Bots work on a cloud computer with its own browser It hands YOU the controls for passwords, 2FA and payments You set what it has to ask before doing HERMES WINS Pick your model. Bring your own key Runs on your laptop or server. Memory stays on your machine Free and open source THE CATCH No model picker. And the weekly usage limits are real. Not technical but still want to customize your agent? Grok Bot is the one. Rules go in the description. Job went well? Say "Save the process we just used as a skill." Want it every Monday? That's a routine. All plain English. No code. You're not paying for the Bot. You're paying to never babysit a server.view source ↗
-
• twitter 1w positiveWhat’s one thing you do to make AI-built websites look less like AI slop? I can get a working first draft quickly. Making it feel designed for that specific client is the hard part. → Visual references? → A design skill? → Better prompts? → Editing the CSS yourself? What has actually made the biggest difference for… What’s one thing you do to make AI-built websites look less like AI slop? I can get a working first draft quickly. Making it feel designed for that specific client is the hard part. → Visual references? → A design skill? → Better prompts? → Editing the CSS yourself? What has actually made the biggest difference for you?view source ↗