•twitter5667
listening to
timeline
5667
mentions
tracked
tracked
net sentiment
+24%
1799 positive
3435 neutral
433 negative
volume & sentiment over time
Sep 30 – Oct 5, 2026 · Daily
sources
most discussed
—
active days
6
mentions
Showing 1676–1700 of 5667 mentions
-
• twitter 1d neutral𝕏 has the people shaping the future What do Instagram and LinkedIn have? 𝕏 has the people shaping the future What do Instagram and LinkedIn have?view source ↗
-
• twitter 1d neutralStarting small is criminally underrated. You don't need: • A huge team. • A giant office. • A massive investment. • A complicated business model. You need enough exposure to the game to develop judgment. Then you increase the size of your bets. Risk should grow with experience. Starting small is criminally underrated. You don't need: • A huge team. • A giant office. • A massive investment. • A complicated business model. You need enough exposure to the game to develop judgment. Then you increase the size of your bets. Risk should grow with experience.view source ↗
-
• twitter 1d neutralHealing doesn’t mean forgetting what happened. It means it no longer gets to control what happens next. Healing doesn’t mean forgetting what happened. It means it no longer gets to control what happens next.view source ↗
-
• twitter 1d neutralBuild quietly. Let the results make the noise. Build quietly. Let the results make the noise.view source ↗
-
• twitter 1d neutralIn August, Koray Kavukcuoglu took the helm of Google DeepMind after Demis Hassabis stepped aside as CEO. In his first media appearance in the role, he sat down at our AI summit with Founder & Editor-in-Chief Jessica Lessin. Watch the exclusive interview here: https://t.co/UxiKQL97gY In August, Koray Kavukcuoglu took the helm of Google DeepMind after Demis Hassabis stepped aside as CEO. In his first media appearance in the role, he sat down at our AI summit with Founder & Editor-in-Chief Jessica Lessin. Watch the exclusive interview here: https://t.co/UxiKQL97gYview source ↗
-
• twitter 1d neutralYour system prompt is 8k tokens and you pay for it on every single request. Prompt caching exists. Most teams have never switched it on. Your system prompt is 8k tokens and you pay for it on every single request. Prompt caching exists. Most teams have never switched it on.view source ↗
-
• twitter 1d neutralPictured: us watching WandaVision for the first time and trying to put the pieces together. https://t.co/sFrbVwfMze Pictured: us watching WandaVision for the first time and trying to put the pieces together. https://t.co/sFrbVwfMzeview source ↗
-
• twitter 1d negativeI made Synara free and open source. And I’m starting to regret it. https://t.co/a3zKi7NLSd I made Synara free and open source. And I’m starting to regret it. https://t.co/a3zKi7NLSdview source ↗
-
• twitter 1d positiveIn the long run, the right opportunities don't just find you, they recognize you. Sometimes you work so hard toward something only to watch it go to someone else. It feels like losing something that was already yours. But keep working. Keep building. Keep becoming someone the right things are drawn to. What's yours … In the long run, the right opportunities don't just find you, they recognize you. Sometimes you work so hard toward something only to watch it go to someone else. It feels like losing something that was already yours. But keep working. Keep building. Keep becoming someone the right things are drawn to. What's yours will find you. It always does.view source ↗
-
• twitter 1d neutralWE ASKED: Who's the most internationally traveled player on the @dallasmavs? 🤔 The Mavs will face the Rockets in the NBA China Games 2026 in Macao, starting October 9 on NBA TV! https://t.co/FehAnZu98a WE ASKED: Who's the most internationally traveled player on the @dallasmavs? 🤔 The Mavs will face the Rockets in the NBA China Games 2026 in Macao, starting October 9 on NBA TV! https://t.co/FehAnZu98aview source ↗
-
• twitter 1d neutralFrom the streets to the pitch! Show off that fancy footwork in @EASPORTSFC 27 on Nintendo Switch 2: https://t.co/JvKZa03cjB https://t.co/3Xo0Q8yDKl From the streets to the pitch! Show off that fancy footwork in @EASPORTSFC 27 on Nintendo Switch 2: https://t.co/JvKZa03cjB https://t.co/3Xo0Q8yDKlview source ↗
-
• twitter 1d neutralLeft hanging 😅 https://t.co/McClgJvgf5 Left hanging 😅 https://t.co/McClgJvgf5view source ↗
-
• twitter 1d neutralsteal this idea. a new type of calorie tracker. AI chatbot. top 1% of this category makes avg. $4m / month. prompt & strategy below ↓ https://t.co/0OlbF0a1k0 steal this idea. a new type of calorie tracker. AI chatbot. top 1% of this category makes avg. $4m / month. prompt & strategy below ↓ https://t.co/0OlbF0a1k0view source ↗
-
• twitter 1d neutrallabs cut prices 50% this week. open models did that. fwiw we're one of the largest buyers of open model tokens on the planet. every cut makes the GOAT plan even more unfair. 🐐 labs cut prices 50% this week. open models did that. fwiw we're one of the largest buyers of open model tokens on the planet. every cut makes the GOAT plan even more unfair. 🐐view source ↗
-
• twitter 1d positiveAndrej Karpathy built an agent that runs AI research autonomously. The whole setup is just 3 files It's called autoresearch. You give Claude or Codex a small LLM training setup and walk away. The loop: > try one idea in https://t.co/293zgMo6Bh > commit > train for exactly 5 minutes > if score improved - keep it. Wor… Andrej Karpathy built an agent that runs AI research autonomously. The whole setup is just 3 files It's called autoresearch. You give Claude or Codex a small LLM training setup and walk away. The loop: > try one idea in https://t.co/293zgMo6Bh > commit > train for exactly 5 minutes > if score improved - keep it. Worse - git reset > log the result, repeat forever The instructions literally tell the agent to never stop and never ask "should I keep going?", because the human might be asleep. You don't touch the Python at all. You only edit program.md, the agent's instructions. Karpathy calls it the code of your research org. There's even a simplicity rule. A tiny gain that adds 20 lines of hacky code gets thrown out. Same score with less code gets kept. I think this pattern works way beyond LLMs. Anything with one number to beat, one file to edit and a fixed time per run can grind overnight like this. 92k stars, MIT. Runs on one NVIDIA GPU, local or rented, with community forks for Mac, Windows and AMD. repo: https://t.co/79UWtuLyk1view source ↗
-
• twitter 1d neutral"The whole city is pumped" 💪 Wakefield Trinity owner Matt Ellis reacts to their Super League Grand Final victory 🌟 https://t.co/KK7AJ0s4I7 "The whole city is pumped" 💪 Wakefield Trinity owner Matt Ellis reacts to their Super League Grand Final victory 🌟 https://t.co/KK7AJ0s4I7view source ↗
-
• twitter 1d positiveI live in Kazakhstan Most of my clients are in the US That's a 12 hour difference But I'm always in touch with them Need edits, a call or just to talk through the project? The time zone never gets in the way I like this work too much https://t.co/issMMMv0bl I live in Kazakhstan Most of my clients are in the US That's a 12 hour difference But I'm always in touch with them Need edits, a call or just to talk through the project? The time zone never gets in the way I like this work too much https://t.co/issMMMv0blview source ↗
-
• twitter 1d positiveHermes BotScreen gives your AI its own desktop to work on. Keep using your screen while your agent handles its task. Here’s what the setup shown in this video includes: → A separate Linux desktop for the agent. → Its own browser, mouse, keyboard, and terminal. → A live view so you can watch it work. → Browser ses… Hermes BotScreen gives your AI its own desktop to work on. Keep using your screen while your agent handles its task. Here’s what the setup shown in this video includes: → A separate Linux desktop for the agent. → Its own browser, mouse, keyboard, and terminal. → A live view so you can watch it work. → Browser sessions with saved logins. → Scheduled tasks for repeat jobs. → Controls to check, start, and stop the instance. You give Hermes an instruction, then inspect what happens on its screen. Start with a simple website visit and summary. Save this video, you’ll know what an agent with its own desktop looks like. Want the SOP? DM me. 💬view source ↗
-
• twitter 1d neutralAntigravity is miles ahead of Codex. It’s not even close. Antigravity is miles ahead of Codex. It’s not even close.view source ↗
-
• twitter 1d neutralSwitching from Github hosted runners to a self-hosted runner (2023 Mac Mini) has reduced my CI workflows from 35 minutes to 5 minutes. Insane. https://t.co/BnRqwLg16a Switching from Github hosted runners to a self-hosted runner (2023 Mac Mini) has reduced my CI workflows from 35 minutes to 5 minutes. Insane. https://t.co/BnRqwLg16aview source ↗
-
• twitter 1d neutralhttps://t.co/7mjLQQHqeA https://t.co/7mjLQQHqeAview source ↗
-
• twitter 1d positiveAn awesome nickname: “the Baklava Bomber”. An awesome nickname: “the Baklava Bomber”.view source ↗
-
• twitter 1d positiveGive it time. Some blessings are worth waiting a little longer for. Give it time. Some blessings are worth waiting a little longer for.view source ↗
-
• twitter 1d positive🚨 Codex Users!! GPT-6 Astra Light is WAY too good for Codex Plus users If you’re on Plus and want to use Astra, we’ve been testing it for almost 1 hour 30 minutes… And the 5-hour window is honestly insane 20% of the 5-hour window still available after that much testing The limit almost doesn’t even feel real 🚨 Codex Users!! GPT-6 Astra Light is WAY too good for Codex Plus users If you’re on Plus and want to use Astra, we’ve been testing it for almost 1 hour 30 minutes… And the 5-hour window is honestly insane 20% of the 5-hour window still available after that much testing The limit almost doesn’t even feel realview source ↗
-
• twitter 1d neutralportfolio with a human touch https://t.co/wP34BC9gae portfolio with a human touch https://t.co/wP34BC9gaeview source ↗