•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 2976–3000 of 4998 mentions
-
• twitter 1w positivelol I need to write about how @polylane does this because ngl ours is better lol I need to write about how @polylane does this because ngl ours is betterview source ↗
-
• twitter 1w positivea german AI model giving me hope? in this economy?? 🇩🇪 europe desperately needs its own competitive, sovereign AI. otherwise we risk: > our best AI researchers leaving to build companies abroad > missing out on the wealth and prosperity created by the AI boom > european businesses becoming dependent on foreign AI pr… a german AI model giving me hope? in this economy?? 🇩🇪 europe desperately needs its own competitive, sovereign AI. otherwise we risk: > our best AI researchers leaving to build companies abroad > missing out on the wealth and prosperity created by the AI boom > european businesses becoming dependent on foreign AI providers > giving washington and beijing leverage over the technology running our economy > becoming a digital colony (by operating on someone else's terms) and with Kolibri, we finally just got one! > it's cheap to run > sovereign (trained from scratch in Europe) > open weights you can download and run on your own hardware > and holds its own against comparable american and chinese open models on benchmarks. it's not frontier. but it's a solid model european businesses can run on their own infrastructure and build into their products and workflows. really hope this is the start of the tide turning in europe.view source ↗
-
• twitter 1w positiveThanks for 400K on YouTube pals 👌 Got recognised at my kids' local park the other day, things are getting mad Thanks for 400K on YouTube pals 👌 Got recognised at my kids' local park the other day, things are getting madview source ↗
-
• twitter 1w positiveThe Great Wall of China 🇨🇳 https://t.co/kIAhXmMorm The Great Wall of China 🇨🇳 https://t.co/kIAhXmMormview source ↗
-
• twitter 1w positiveYou can now steer Cursor SDK agents while they run. `run.steer()` adds your message to the next turn. If a subagent is mid-task, it moves to the background and keeps working. Cursor posted the change today, and the TypeScript SDK docs cover steering a run in flight. https://t.co/Fr7KrUIrbZ You can now steer Cursor SDK agents while they run. `run.steer()` adds your message to the next turn. If a subagent is mid-task, it moves to the background and keeps working. Cursor posted the change today, and the TypeScript SDK docs cover steering a run in flight. https://t.co/Fr7KrUIrbZview source ↗
-
• twitter 1w positiveFree web search just reached more Hermes users.🚀 A fix on Hermes main lets anonymous Nous guests use managed Perplexity Fast Search without a separate Perplexity API key. Try this: “Find three open-source PDF-to-text tools. Compare scanned-PDF support, local use and licenses. Link the original repos. Flag anything un… Free web search just reached more Hermes users.🚀 A fix on Hermes main lets anonymous Nous guests use managed Perplexity Fast Search without a separate Perplexity API key. Try this: “Find three open-source PDF-to-text tools. Compare scanned-PDF support, local use and licenses. Link the original repos. Flag anything unverified.” Build a shortlist before installing anything. Free applies to managed Fast Search. Model usage and page extraction are separate. Existing provider settings or direct keys can take priority. The guest-access fix isn’t in the latest stable release yet.🤟view source ↗
-
• twitter 1w positivei love this, it is very much similar to the claude code's project-level memories (except the swarm part), most models use it well, gpt in codex is excellent at using and maintaining it. you can set it up using hooks using this repo i made :) https://t.co/ye0TsztxT6 i love this, it is very much similar to the claude code's project-level memories (except the swarm part), most models use it well, gpt in codex is excellent at using and maintaining it. you can set it up using hooks using this repo i made :) https://t.co/ye0TsztxT6view source ↗
-
• twitter 1w positiveHighly recommend adding a 𝗰𝗮𝗰𝗵𝗲 𝗰𝗼𝘂𝗻𝘁𝗱𝗼𝘄𝗻 to your Claude Code status line. Simply put, every time you send a message, Claude Code sends the whole conversation to the model again. The cache lets it skip reprocessing everything that came before, so replies come back faster and count less against your usage limits. But t… Highly recommend adding a 𝗰𝗮𝗰𝗵𝗲 𝗰𝗼𝘂𝗻𝘁𝗱𝗼𝘄𝗻 to your Claude Code status line. Simply put, every time you send a message, Claude Code sends the whole conversation to the model again. The cache lets it skip reprocessing everything that came before, so replies come back faster and count less against your usage limits. But the cache expires, and the first message after that makes it process 𝘁𝗵𝗲 𝘄𝗵𝗼𝗹𝗲 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻 𝗳𝗿𝗼𝗺 𝘀𝗰𝗿𝗮𝘁𝗰𝗵. Normally you can't see how much time it has left. With this, your status line always shows the minutes left, and once the cache expires, it tells you how many tokens your next message will re-cache. If it's about to run out and you still have something to ask, ask it now. The official docs have a few more things I didn't know: → On a subscription, within your plan's usage, the main conversation's cache lasts 1 hour, while subagents and workflows get 5 minutes. On an API key, or once you're past your limit and on usage credits, the main conversation gets 5 minutes too → Switching models mid-session recomputes everything, and with opusplan, every move into or out of plan mode counts as a switch → Editing CLAUDE.md mid-session doesn't break the cache, but the edit doesn't apply either until /clear, /compact or a restart → Went down the wrong path and want a redo? /rewind back and it can pick up the cache you already had; /compact builds a new one Send this prompt and the link below to Opus 5.5 👇 "Read this doc, plus the status line field reference it links to, and add a cache segment to my Claude Code status line. 1. If I already have a status line, don't replace it. Save stdin first so the original command and the cache segment read the same input, leave the original output as is, and put the cache segment on its own line after it. If I don't have one, create one. 2. Get the data from prompt_cache in the input: warm, ttl, expires_at, hit_ratio, misses, last_miss_cause, recache_tokens_if_cold. Show nothing until prompt_cache appears, and skip any fiview source ↗
-
• twitter 1w positiveA good Claude mods built on top of Jev + @treg_ai A good Claude mods built on top of Jev + @treg_aiview source ↗
-
• twitter 1w positivewe're hiring our first engineer at @heardlabshq 🎧 + $1k referral bonus you: • AI/ML background • have shipped agents that actually work in prod • great product taste how to apply: comment or DM with your background, or try the product and send us your top 3 pieces of feedback. that's it. $1k to whoever refers the hi… we're hiring our first engineer at @heardlabshq 🎧 + $1k referral bonus you: • AI/ML background • have shipped agents that actually work in prod • great product taste how to apply: comment or DM with your background, or try the product and send us your top 3 pieces of feedback. that's it. $1k to whoever refers the hire 👇 https://t.co/wJpphsLw0Zview source ↗
-
• twitter 1w positiveI built a Claude skill that runs underneath everything I write. Today it's yours, free. It catches the patterns that make copy sound like AI wrote it, then fixes them. 25+ checks, including the new ones. The Economist analyzed 1.2 million words of AI text this summer, and the models barely use em dashes anymore. Now t… I built a Claude skill that runs underneath everything I write. Today it's yours, free. It catches the patterns that make copy sound like AI wrote it, then fixes them. 25+ checks, including the new ones. The Economist analyzed 1.2 million words of AI text this summer, and the models barely use em dashes anymore. Now they under-punctuate: fewer commas, almost no parentheses, long smooth sentences that never stop for breath. It started as a json checklist 800+ of you downloaded. Install it once and forget about it. Open source: https://t.co/vA5bcSBNlTview source ↗
-
• twitter 1w positiveThanks to AI coding, we will soon have free (open source) versions of Photoshop and all the other expensive Adobe tools! This is just the beginning. AI coding is going unlock so many cool things that even the AI haters are going to be delighted! Thanks to AI coding, we will soon have free (open source) versions of Photoshop and all the other expensive Adobe tools! This is just the beginning. AI coding is going unlock so many cool things that even the AI haters are going to be delighted!view source ↗
-
• twitter 1w positiveNeon backend 🫶 @tan_stack We built Atlas: a photo library you can search by text, image, or face The stack: - Lakebase Postgres for metadata, captions, and embeddings - Object Storage for the photos - Managed Better Auth for sign-in - Lakebase Search to rank embeddings inside Postgres - TanStack Start for the app an… Neon backend 🫶 @tan_stack We built Atlas: a photo library you can search by text, image, or face The stack: - Lakebase Postgres for metadata, captions, and embeddings - Object Storage for the photos - Managed Better Auth for sign-in - Lakebase Search to rank embeddings inside Postgres - TanStack Start for the app and server functionsview source ↗
-
• twitter 1w positive"Give it to me, come up, hit, and roll again." THIS ALLEY-OOP PAIRING 🔥 LaMelo Ball and Rudy Gobert building their lob connection during practice! https://t.co/OveWk1UhQC "Give it to me, come up, hit, and roll again." THIS ALLEY-OOP PAIRING 🔥 LaMelo Ball and Rudy Gobert building their lob connection during practice! https://t.co/OveWk1UhQCview source ↗
-
• twitter 1w positiveI just bought the new MacBook Air 🤩 What should I install first? https://t.co/cqemgMXKoj I just bought the new MacBook Air 🤩 What should I install first? https://t.co/cqemgMXKojview source ↗
-
• twitter 1w positiveimo one of the underhyped launches of @OpenAI dev day were all the rad updates to Sites, especially connected data + site-as-infra which is why it was so fun to sit down with @simpsoka (product lead) and talk about the practical + fun ways she uses sites. in this live ep we cover: - how sites can use your connected e… imo one of the underhyped launches of @OpenAI dev day were all the rad updates to Sites, especially connected data + site-as-infra which is why it was so fun to sit down with @simpsoka (product lead) and talk about the practical + fun ways she uses sites. in this live ep we cover: - how sites can use your connected enterprise data + user permissions to create micro SaaS everyone can use - fun consumer use cases, like a Spotify playlist sans baby music - the internals of Sites + how it can host so much more than just webpages anyone struggling with "where the f do we deploy all these internal tools" should pay attention 👇 Full ep on YT: https://t.co/52I4wPnxdyview source ↗
-
• twitter 1w positiveplease if you're talented and ambitious do not work on saas. pick a physical problem in the world and take it on, the models are too good not to please if you're talented and ambitious do not work on saas. pick a physical problem in the world and take it on, the models are too good not toview source ↗
-
• twitter 1w positiveOCR is dead 🪦 Long live agentic OCR! Traditional OCR makes one pass and hands back whatever text it got. Tables get flattened, charts disappear, and multi-column layouts come out scrambled. Nothing checks the output. Agentic OCR treats parsing as a loop instead: ✅ layout-aware reading order ✅ smart routing of hard e… OCR is dead 🪦 Long live agentic OCR! Traditional OCR makes one pass and hands back whatever text it got. Tables get flattened, charts disappear, and multi-column layouts come out scrambled. Nothing checks the output. Agentic OCR treats parsing as a loop instead: ✅ layout-aware reading order ✅ smart routing of hard elements to the right model ✅ multi-pass verification and self-correction ✅ multimodal parsing of charts, images, and complex tables Logan Markewich, LlamaIndex's Head of Open Source, wrote about what that shift looks like in practice and where it's still struggling. Link to the breakdown in the comments below.view source ↗
-
• twitter 1w positiveif anyone wants to play ball in toronto, jared's awesome if anyone wants to play ball in toronto, jared's awesomeview source ↗
-
• twitter 1w positivegyotaku runs on windows now. one line in powershell, same instant search, fully offline. your screenshots never leave your machine. gyotaku runs on windows now. one line in powershell, same instant search, fully offline. your screenshots never leave your machine.view source ↗
-
• twitter 1w positiveCORRUPTION: The same dead people voting for Democrats have been collecting almost $2 billion from Obamacare. Dr. Oz just terminated them for good. CORRUPTION: The same dead people voting for Democrats have been collecting almost $2 billion from Obamacare. Dr. Oz just terminated them for good.view source ↗
-
• twitter 1w positiveWell, Computer Use in Codex *used* to be a bundled skill which users could modify (and configure their desired guardrails, accordingly). On quick review, looks like latest Codex releases removed this and guardrails are baked into compiled/minified code 😔 https://t.co/sR3bXfM01B Well, Computer Use in Codex *used* to be a bundled skill which users could modify (and configure their desired guardrails, accordingly). On quick review, looks like latest Codex releases removed this and guardrails are baked into compiled/minified code 😔 https://t.co/sR3bXfM01Bview source ↗
-
• twitter 1w positiveHow I wish every Indian student should listen to him found him back in my uni days Went to US in a conference (fully sponsored) Contributed very minutely to open source - life had other plans Leading the complete AI engineering for my company right now (I am 27) The second order effects of doing things early in life… How I wish every Indian student should listen to him found him back in my uni days Went to US in a conference (fully sponsored) Contributed very minutely to open source - life had other plans Leading the complete AI engineering for my company right now (I am 27) The second order effects of doing things early in life are crazy @kunalstwt thanks for your content manview source ↗
-
• twitter 1w positiveRT @RhysSullivan: t3 code is so much better than the labs products its not even close RT @RhysSullivan: t3 code is so much better than the labs products its not even closeview source ↗
-
• twitter 1w positiveThe hardware is clearly meant to be a loss leader. $3.5k is at or below the market price of the parts, before you even account for labor. Profitable hardware companies hit 30-50% margins, which would be $5-7k a box here (which no one would ever pay for these specs) Based on the use cases in the launch video, this is a… The hardware is clearly meant to be a loss leader. $3.5k is at or below the market price of the parts, before you even account for labor. Profitable hardware companies hit 30-50% margins, which would be $5-7k a box here (which no one would ever pay for these specs) Based on the use cases in the launch video, this is a local Instinct / Muse play. It will have hyper contextual data on you, your health, your interests and be able to recommend the best products to buy. And Ghost will take a cut of that. Agentic commerce is a much more attractive market than reselling Nvidia chips You can't assemble this spec yourself for $3.5k right now, so purely as parts it's a decent deal. You can also plug in another agent like OpenCode. But they obviously want you to buy into their agent and ecosystem, so I would be concerned about 3rd party agents staying 1st classview source ↗