•twitter6204
listening to
timeline
6204
mentions
tracked
tracked
net sentiment
+24%
1974 positive
3760 neutral
470 negative
volume & sentiment over time
Sep 30 – Oct 5, 2026 · Daily
sources
most discussed
—
active days
6
mentions
Showing 1251–1275 of 6204 mentions
-
• twitter 18h positiveA new AI model just topped real-world coding benchmarks. Not a rumor, not a leak. Google announced it. Meet Gemini 4 Argon, released September 30, 2026. 🚀 It's built for "long-horizon workflows": tasks with lots of steps, lots of thinking, and lots of back-and-forth. → 77.9% on DeepSWE v1.1, a new state of the art… A new AI model just topped real-world coding benchmarks. Not a rumor, not a leak. Google announced it. Meet Gemini 4 Argon, released September 30, 2026. 🚀 It's built for "long-horizon workflows": tasks with lots of steps, lots of thinking, and lots of back-and-forth. → 77.9% on DeepSWE v1.1, a new state of the art for real-world software engineering → Up to 1M output tokens in one response (it was 64K before) → Multimodal reasoning across charts, videos, and documents → Agentic, so it can run code and take actions, not just answer questions Google thousands of employees are already using it for coding, deep research, and document writing. Save this video, you'll know exactly why this launch matters before your feed explains it for you. Want the SOP? DM me.view source ↗
-
• twitter 18h neutralJapan service sector activity slows from 5-month high, PMI shows https://t.co/DF1Rl1n6nS https://t.co/DF1Rl1n6nS Japan service sector activity slows from 5-month high, PMI shows https://t.co/DF1Rl1n6nS https://t.co/DF1Rl1n6nSview source ↗
-
• twitter 18h positive30 things that will get you hired as an AI engineer in 2026: – Ship one agent real people use every week – Know when NOT to use an LLM – Write evals before writing prompts – Build an MCP server from scratch – Connect an agent to 5+ real tools – Design memory that forgets the right things – Pick the cheapest mod… 30 things that will get you hired as an AI engineer in 2026: – Ship one agent real people use every week – Know when NOT to use an LLM – Write evals before writing prompts – Build an MCP server from scratch – Connect an agent to 5+ real tools – Design memory that forgets the right things – Pick the cheapest model that passes your evals – Route easy tasks to small models, hard ones to big ones – Cut your inference bill in half and show the numbers – Cache repeat answers so they cost almost nothing – Make long workflows resume after a crash – Validate every model output against a schema – Sandbox every tool call – Defend against prompt injection from docs and web pages – Keep secrets and PII out of the context window – Add human approval before risky actions – Trace every model call and tool call – Debug a failing agent from its logs alone – Grade the agent's steps, not just its final answer – Turn real user failures into test cases – Block deploys when eval scores drop – Test new models on a slice of real traffic first – Add a fallback when your model provider goes down – Stream responses so users never wait on a spinner – Fine-tune a small model on your best outputs – Version prompts like code – Measure cost and latency per task – Write a README anyone understands in 5 minutes – Record a 60-second demo that shows it working – Post what broke and how you fixed it Nobody asks for your certificate. They ask "what did you ship?" Bookmark this + send it to someone prepping for AI interviews.view source ↗
-
• twitter 18h neutralKeep going. Some of your happiest memories are still waiting to be made. Keep going. Some of your happiest memories are still waiting to be made.view source ↗
-
• twitter 18h neutralI worked so hard to have my wife approve my $100//mo. budget. I need to use it wisely Codex Pro or Claude Code Max, tell me I worked so hard to have my wife approve my $100//mo. budget. I need to use it wisely Codex Pro or Claude Code Max, tell meview source ↗
-
• twitter 18h neutralwtf is a software factory wtf is a software factoryview source ↗
-
• twitter 18h neutralIf you're building an agentic system Use idempotency around the actions. Put caching around model calls and tool results. these two alone stop a lot of the wasted time and tokens in a flow If you're building an agentic system Use idempotency around the actions. Put caching around model calls and tool results. these two alone stop a lot of the wasted time and tokens in a flowview source ↗
-
• twitter 18h neutralah shht mailflare not found help me find it 😂 👉 https://t.co/WNeQigKRjQ https://t.co/AVOoHMsdOH ah shht mailflare not found help me find it 😂 👉 https://t.co/WNeQigKRjQ https://t.co/AVOoHMsdOHview source ↗
-
• twitter 18h neutralDesigned for Property Decisions https://t.co/7Rog3FnXPf Designed for Property Decisions https://t.co/7Rog3FnXPfview source ↗
-
• twitter 18h neutralUS military says it struck vessel in Caribbean, killing four https://t.co/SdBSN4E57Z https://t.co/SdBSN4E57Z US military says it struck vessel in Caribbean, killing four https://t.co/SdBSN4E57Z https://t.co/SdBSN4E57Zview source ↗
-
• twitter 18h neutral@NarukeAlpha It will be the furthest possible opposite of Omarchy. Chatted w/ a maintainer about it last week :) @NarukeAlpha It will be the furthest possible opposite of Omarchy. Chatted w/ a maintainer about it last week :)view source ↗
-
• twitter 18h positiveRT @dabit3: strong people lift others up. weak people pull others down. winners focus on winning. losers focus on winners. people doing gre… RT @dabit3: strong people lift others up. weak people pull others down. winners focus on winning. losers focus on winners. people doing gre…view source ↗
-
• twitter 18h positiveWow this is amazing 🔥 Google is now including Colab inside their Gemini plans... so you have access to premium GPUs!! You can basically: - Run any open source model - Even image/audio/video gen - Fine-tune a lot of OSS models Everything on an expensive Nvidia GPU without having to rent it or own it. So many possibi… Wow this is amazing 🔥 Google is now including Colab inside their Gemini plans... so you have access to premium GPUs!! You can basically: - Run any open source model - Even image/audio/video gen - Fine-tune a lot of OSS models Everything on an expensive Nvidia GPU without having to rent it or own it. So many possibilities. Time to build!view source ↗
-
• twitter 18h positive@maria_rcks My experience with grok in my Tesla could be much better as well lol @maria_rcks My experience with grok in my Tesla could be much better as well lolview source ↗
-
• twitter 18h positiveToday is National Be Nice Day. Today is National Be Nice Day.view source ↗
-
• twitter 18h neutralProduct strategy starts long before the first wireframe. Before opening Figma: What problem are we solving? What evidence do we have? Which assumptions are we making? Why does this matter to the business? Wireframes help define the solution. Strategy helps determine whether it's worth building. #ProductStrategy #B2B… Product strategy starts long before the first wireframe. Before opening Figma: What problem are we solving? What evidence do we have? Which assumptions are we making? Why does this matter to the business? Wireframes help define the solution. Strategy helps determine whether it's worth building. #ProductStrategy #B2BSaaSview source ↗
-
• twitter 18h neutralRT @AlexFinn: If you’re using Grok Bot, Dots, Muse, or Hermes, your agent is capable of WAY more than you think You’re probably only getti… RT @AlexFinn: If you’re using Grok Bot, Dots, Muse, or Hermes, your agent is capable of WAY more than you think You’re probably only getti…view source ↗
-
• twitter 18h positiveWait what, Codex users get a free reset every day for 28 days if OpenAI doesn’t ship a meaningful update? Wait what, Codex users get a free reset every day for 28 days if OpenAI doesn’t ship a meaningful update?view source ↗
-
• twitter 19h positiveBetting on this being better than omarchy. Betting on this being better than omarchy.view source ↗
-
• twitter 19h neutral@aAhmidAziz @NousResearch Then fix it! @aAhmidAziz @NousResearch Then fix it!view source ↗
-
• twitter 19h neutralBuilt a SaaS as a hobby. Deployed it. Never promoted it. 4 months later: → 160k visitors → 1,239 signups The surprising part: visitors from ChatGPT convert at 28%, compared to 3% from Google. Maybe this hobby deserves more attention 👀 #BuildInPublic https://t.co/rVbSHeKvbA Built a SaaS as a hobby. Deployed it. Never promoted it. 4 months later: → 160k visitors → 1,239 signups The surprising part: visitors from ChatGPT convert at 28%, compared to 3% from Google. Maybe this hobby deserves more attention 👀 #BuildInPublic https://t.co/rVbSHeKvbAview source ↗
-
• twitter 19h positiveWow, @steipete delivers a new-to-me YOLO/skip permissions gem for computer use: What it unlocks: "per-app limits get lifted. Today those stop me from clicking or typing in browsers and from typing in terminals and IDEs. That would let me type into Chrome or cmux directly" Wow, @steipete delivers a new-to-me YOLO/skip permissions gem for computer use: What it unlocks: "per-app limits get lifted. Today those stop me from clicking or typing in browsers and from typing in terminals and IDEs. That would let me type into Chrome or cmux directly"view source ↗
-
• twitter 19h neutralLooking back, it is so interesting that we all rushed in 2023-24 to learn LangChain and many other frameworks. And now, almost all major coding harnesses and agents are just hand-rolled loops. Turns out, the abstraction was the overhead. Looking back, it is so interesting that we all rushed in 2023-24 to learn LangChain and many other frameworks. And now, almost all major coding harnesses and agents are just hand-rolled loops. Turns out, the abstraction was the overhead.view source ↗
-
• twitter 19h neutralSupporters of jailed ex-PM Imran Khan march to Pakistan capital https://t.co/H2xACscQRr Supporters of jailed ex-PM Imran Khan march to Pakistan capital https://t.co/H2xACscQRrview source ↗
-
• twitter 19h neutrali was in a random bookstore & came across a book that described achieving moksha as achieving “super intelligence”. the hindu religion figured it out a while ago. https://t.co/bY8QvTpi3v i was in a random bookstore & came across a book that described achieving moksha as achieving “super intelligence”. the hindu religion figured it out a while ago. https://t.co/bY8QvTpi3vview source ↗