•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 3376–3400 of 4998 mentions
-
• twitter 1w positiveAll systems are looking good, and weather is favorable for liftoff. Propellant load is about to begin for Falcon 9’s launch of the @SemperCitiusSDA's fourth Tranche 1 data transport mission from California → https://t.co/wz4JbJVy1b https://t.co/Bqj0Ksy8NJ All systems are looking good, and weather is favorable for liftoff. Propellant load is about to begin for Falcon 9’s launch of the @SemperCitiusSDA's fourth Tranche 1 data transport mission from California → https://t.co/wz4JbJVy1b https://t.co/Bqj0Ksy8NJview source ↗
-
• twitter 1w positiveIf you're on Opus 5.5, read this one. The Claude team wrote a guide on getting the most out of Opus 5.5. It works longer on its own and thinks before every reply, so your prompts need to change. The tips I'd start with: • Give the whole task in one message. Say what "done" means and when it should stop and ask. • … If you're on Opus 5.5, read this one. The Claude team wrote a guide on getting the most out of Opus 5.5. It works longer on its own and thinks before every reply, so your prompts need to change. The tips I'd start with: • Give the whole task in one message. Say what "done" means and when it should stop and ask. • Delete "think carefully" lines. It thinks anyway, and replies start sooner without them. • Add a stop rule to CLAUDE.md: keep going unless it's blocked or about to do something destructive. • Keep the task list in a file so it survives context compaction. • Split codebase-wide audits, migrations, and reviews across subagents, then check each one's evidence. Read the full guide before your next long run. https://t.co/XJHxpFtdpfview source ↗
-
• twitter 1w positiveThis is the ChatGPT moment for Voice AI. A ten-second sample can show that a synthetic voice sounds natural. But it tells little about whether that voice can stay coherent across an entire conversation. Real callers ask follow-up questions, change their requirements, and interrupt before the agent finishes speaking. … This is the ChatGPT moment for Voice AI. A ten-second sample can show that a synthetic voice sounds natural. But it tells little about whether that voice can stay coherent across an entire conversation. Real callers ask follow-up questions, change their requirements, and interrupt before the agent finishes speaking. The voice must respond quickly without resetting its tone and pacing after every turn. This creates two memory problems. - The LLM needs to remember what the caller said. - The speech model needs to retain how the agent has been speaking. A transcript handles the first problem, but not the second. It preserves the words, not their delivery or how much of the previous response the caller actually heard. Deepgram's Flux TTS handles this by maintaining acoustic state inside a persistent streaming session. Each response begins with the speech history from earlier turns, and the first audio arrives in under 200 ms. Flux also handles interruptions through playback tracking. The application sends the point where playback stopped, and the API reports which words the caller heard. That gives the LLM an accurate conversation history for its next response. In a blind listening evaluation across roughly 14,400 paired comparisons, Flux's best voice recorded a 73.4% naturalness win rate. Cartesia Sonic 3.5 recorded 72.0%. Build with Flux TTS here: https://t.co/CT2TXyjZ1Y I recorded a conversation with a car dealership agent about hybrid and electric cars. I changed my requirements and asked follow-up questions to hear how the voice behaved across the full exchange. The recording is below, and I worked with the Deepgram team on this post to put it together.view source ↗
-
• twitter 1w positiveToday, I am giving a live training on how to build an audience in 2026. It’s at 4 pm BST. It’s free to join. Just reply ‘training’ and we will send the room link. Today, I am giving a live training on how to build an audience in 2026. It’s at 4 pm BST. It’s free to join. Just reply ‘training’ and we will send the room link.view source ↗
-
• twitter 1w positivesomeone asked what deploying AI agents at a Fortune 100 actually looks like: https://t.co/ioyuJXeHWc someone asked what deploying AI agents at a Fortune 100 actually looks like: https://t.co/ioyuJXeHWcview source ↗
-
• twitter 1w positiveOpenAI gave me one month of Plus($20) plan for free. because i wanted to cancel my subscription https://t.co/ovGPN6s64w OpenAI gave me one month of Plus($20) plan for free. because i wanted to cancel my subscription https://t.co/ovGPN6s64wview source ↗
-
• twitter 1w positiveI’m increasingly convinced OpenAI is ~6 months ahead of Anthropic internally. The mistake people make is comparing what’s public and assuming that’s the frontier. OpenAI has every incentive to sit on its best models until it actually needs them. If you already have the lead, why burn your compute advantage and reveal… I’m increasingly convinced OpenAI is ~6 months ahead of Anthropic internally. The mistake people make is comparing what’s public and assuming that’s the frontier. OpenAI has every incentive to sit on its best models until it actually needs them. If you already have the lead, why burn your compute advantage and reveal your hand? Anthropic ships something great. OpenAI waits. Then drops the next tier. I don’t think we’re seeing OpenAI’s frontier. We’re seeing what they’re comfortable releasing.view source ↗
-
• twitter 1w positiveSIMPLE SELLS because SIMPLE SCALES. Honestly, I prefer the UI version. Terminals are dead. Thanks for the amazing lesson, Codex, Antigravity, and Devin. You can't customise a terminal enough. Not saying that Linux doesn't have fans, but as I am getting older, I appreciate Mac's simplicity. SIMPLE SELLS because SIMPLE SCALES. Honestly, I prefer the UI version. Terminals are dead. Thanks for the amazing lesson, Codex, Antigravity, and Devin. You can't customise a terminal enough. Not saying that Linux doesn't have fans, but as I am getting older, I appreciate Mac's simplicity.view source ↗
-
• twitter 1w positiveGreat post by @hezarfendd about using AI for learning new things https://t.co/0fj94i34cS Great post by @hezarfendd about using AI for learning new things https://t.co/0fj94i34cSview source ↗
-
• twitter 1w positiveGun to your head Most beautiful website you’ve ever seen Go Gun to your head Most beautiful website you’ve ever seen Goview source ↗
-
• twitter 1w positiveA 125 billion parameter model now runs on a 12 GB gaming card. At 94 tokens a second. No server. No API bill. Nothing leaves your PC. It's called Strata, by Niko1221. Free, MIT licence, 11.8K GitHub stars in 11 days. The model is Qwen3.8-Flash-Next. Models this size normally live on servers with hundreds of gigabyte… A 125 billion parameter model now runs on a 12 GB gaming card. At 94 tokens a second. No server. No API bill. Nothing leaves your PC. It's called Strata, by Niko1221. Free, MIT licence, 11.8K GitHub stars in 11 days. The model is Qwen3.8-Flash-Next. Models this size normally live on servers with hundreds of gigabytes of GPU memory. How a 12 GB card pulls it off: The model is 24,576 small experts, and each word only uses 10 of them. Strata keeps the busiest few thousand on the GPU, all of them in RAM, and a 29 GB lookup table on your SSD. The CPU works on the rest at the same time. Then a small helper guesses the next few words and the big model checks them in one go. Same answer, 1.6 to 1.8x sooner. Their numbers on an RTX 5070 (12 GB) with 64 GB RAM: 94 tokens/s on a short chat 76 tokens/s at 128K context 2,650 tokens/s reading a 32K prompt On an AMD RX 9070 XT: 60 tokens/s. Setup is one file. Windows: double-click START-HERE.bat. Linux: ./setup.sh. Or paste one line into Claude Code or Codex and it sets itself up. It serves OpenAI and Anthropic style APIs on localhost:8080, so Hermes, Claude Code and Codex can point straight at it. The catch, because there is one: You need 32 to 64 GB of RAM, not just the card. About 80 GB of disk. The fastest size is 2-bit, so you trade some quality for speed. And your PC can freeze for 1 to 3 minutes while it loads the first time. Still. This used to mean renting a server. https://t.co/vWVp5BC13eview source ↗
-
• twitter 1w positiveImages + prompts 🔥 Just tried this out yesterday 😂 Creativity isn’t limited to development https://t.co/pGgc6jkyza Images + prompts 🔥 Just tried this out yesterday 😂 Creativity isn’t limited to development https://t.co/pGgc6jkyzaview source ↗
-
• twitter 1w positiveWhat did Hermes take off your plate this week? Research, fixing code, sorting files—or something else? Share one task it finished and how you checked the result. Example: “Sorted my invoices. I checked the dates and totals against the originals.” Still looking for your first useful win? Tell us the task you want he… What did Hermes take off your plate this week? Research, fixing code, sorting files—or something else? Share one task it finished and how you checked the result. Example: “Sorted my invoices. I checked the dates and totals against the originals.” Still looking for your first useful win? Tell us the task you want help with.view source ↗
-
• twitter 1w positiveWHO CAN ACTUALLY GET A DOT RIGHT NOW dots launched on September 29, but not for everyone. the access map this week: > ChatGPT Pro, from $100 a month: first dot included, outside the EEA, Switzerland and the UK for now > Business Premium: rolling out in all supported regions, first dot included > Enterprise, Edu and H… WHO CAN ACTUALLY GET A DOT RIGHT NOW dots launched on September 29, but not for everyone. the access map this week: > ChatGPT Pro, from $100 a month: first dot included, outside the EEA, Switzerland and the UK for now > Business Premium: rolling out in all supported regions, first dot included > Enterprise, Edu and Healthcare: a beta that stays off until a workspace admin turns it on > Free, Go and Plus: no dot and where you set one up: > the ChatGPT desktop app or a desktop browser. you cannot create a dot on mobile, and mobile web is not supported > after that you can talk to it from the mobile app, Slack and Teams > texting is a limited beta, Pro users in the US only > it cannot call you, and it has no email address of its own. you can connect your personal email the rollout is gradual even on an eligible plan it can take a few days to show up, so check the list before you upgrade a plan for thisview source ↗
-
• twitter 1w positiveOpen source is f***ing cool. Completely blows my mind that anyone can adjust the product with some prompts. Sebastian added the kanban. Open source is f***ing cool. Completely blows my mind that anyone can adjust the product with some prompts. Sebastian added the kanban.view source ↗
-
• twitter 1w positiveCodex /usage analytics view looks cool https://t.co/gNnEK9lXLI Codex /usage analytics view looks cool https://t.co/gNnEK9lXLIview source ↗
-
• twitter 1w positive🔥 New Slot Spotlight: Spice (NetEnt) Looking for a fresh slot to try? 🌶 Spice by NetEnt • Provider: NetEnt • Volatility: Medium-High • RTP: 96.09% • Max Win: 10,000x your bet Features: - Free Spins Bonus - Expanding Wilds - Multipliers - Sticky Wild mechanics Perfect for players who enjoy balanced gameplay with exci… 🔥 New Slot Spotlight: Spice (NetEnt) Looking for a fresh slot to try? 🌶 Spice by NetEnt • Provider: NetEnt • Volatility: Medium-High • RTP: 96.09% • Max Win: 10,000x your bet Features: - Free Spins Bonus - Expanding Wilds - Multipliers - Sticky Wild mechanics Perfect for players who enjoy balanced gameplay with exciting bonus rounds. Read the full review: https://t.co/0FxOmwvIKc 🎮 Stay tuned for more updates! #NetEnt #Slots #OnlineCasino #CryptoCasino #FullhouseGGview source ↗
-
• twitter 1w positiveWe’re hiring a Fullstack Engineer. Love building from 0→1 and shipping fast? You might be our person. What we’re looking for: • Strong fullstack development experience • Good product sense + ownership • Comfortable in a fast-paced environment • Idea -build- ship. 🚀 https://t.co/kRY9UWnALD We’re hiring a Fullstack Engineer. Love building from 0→1 and shipping fast? You might be our person. What we’re looking for: • Strong fullstack development experience • Good product sense + ownership • Comfortable in a fast-paced environment • Idea -build- ship. 🚀 https://t.co/kRY9UWnALDview source ↗
-
• twitter 1w positiveOne great surprise about posting online is finding how many like minded people are out there. That weird idea you have that makes you feel alone in the real world? It will resonate with some stranger you've never met. The only way to find out is to post it. One great surprise about posting online is finding how many like minded people are out there. That weird idea you have that makes you feel alone in the real world? It will resonate with some stranger you've never met. The only way to find out is to post it.view source ↗
-
• twitter 1w positiveIt is extraordinary that India, still a relatively poor country, can achieve something as selfless as protecting other creatures. We explain how it became remarkably good at animal conservation https://t.co/TP0ik9s4AT It is extraordinary that India, still a relatively poor country, can achieve something as selfless as protecting other creatures. We explain how it became remarkably good at animal conservation https://t.co/TP0ik9s4ATview source ↗
-
• twitter 1w positiveA VIBE CODER’S NIGHTMARE: the product works, the code is clean, and not one person opens it. A VIBE CODER’S NIGHTMARE: the product works, the code is clean, and not one person opens it.view source ↗
-
• twitter 1w positiveThis week, let's win again! Happy new week 🤍 https://t.co/JrSTsO7W1L This week, let's win again! Happy new week 🤍 https://t.co/JrSTsO7W1Lview source ↗
-
• twitter 1w positivethis is free f*cking gold for anyone building a second brain his second brain article pulled 8 million views, so he packed the whole system into one repo the repo, the guide, the tools, the learning path, all free • the guide > 10 sections, 65 pages, from the core idea to troubleshooting > 5 tracks on top, 44 pages,… this is free f*cking gold for anyone building a second brain his second brain article pulled 8 million views, so he packed the whole system into one repo the repo, the guide, the tools, the learning path, all free • the guide > 10 sections, 65 pages, from the core idea to troubleshooting > 5 tracks on top, 44 pages, 15 of them build guides with working code • the machine (.claude/) > 18 agent skills, one per workflow > 72 slash commands - /ingest-pdf, /ingest-youtube, /ingest-voice, /backfill > 6 subagents - curator, linker, researcher, reviewer, ingestor, graph-analyst > 4 are read-only, so nothing rewrites your vault behind your back • the scripts (plain Python, zero dependencies) > graph export, link checker, vault stats, chat converter, site builder • the starter vault > its own CLAUDE.md with page contracts and linking rules > raw/ stays untouched once it lands, wiki/ is what the agent keeps up > log.md - one line per run, so every change is traceable • 87 vetted resources > 28 tools, 26 Obsidian plugins, 15 repos, 12 skills, plus papers and articles five tracks to choose from: > knowledge graphs > Jev engineering > agent harnesses > loop engineering > eval engineering new to this? start with the second brain guide. already deep in it? jump straight to the tracks the Jev engineering and loop engineering tracks alone are worth the clone a consultant bills four figures to build you a research system like this. here it's a public repo under MIT ↳ https://t.co/3zyoyQy0afview source ↗
-
• twitter 1w positiveFree AI credits Grok Bot is ready for download This bot informs you of posts in X that give away free credits Try it, I already found a few! 😁 https://t.co/valuMg2ej3 https://t.co/dlHL3fnI9X Free AI credits Grok Bot is ready for download This bot informs you of posts in X that give away free credits Try it, I already found a few! 😁 https://t.co/valuMg2ej3 https://t.co/dlHL3fnI9Xview source ↗
-
• twitter 1w positiveWhat sort of agent plugins/tools etc. are you using to get the most of out of local models in agents? I've been experimenting with this but I would love to hear community insights How can you push that Qwen 27B model to punch above its weight without altering the model iteself? What sort of agent plugins/tools etc. are you using to get the most of out of local models in agents? I've been experimenting with this but I would love to hear community insights How can you push that Qwen 27B model to punch above its weight without altering the model iteself?view source ↗