LLMs Achieve Rapid Logistic Gains on New Benchmark
Exponential improvements* everywhere for those with the eyes to see them. This is a cool benchmark, and was impossible for early non-reasoner LLMs to do at all. * Okay, technically "logistic improvement" because the maximum score is bounded at 100 (and logistic has a lower AIC) https://t.co/9kVj4o7Gz0
Specialized Compute Models Offer Cheaper Attention in LLMs
This is kind of insane LLM research. It postulates a world of specialized compute models with cheaper attention embedded with an LLMs weights.
Editing Live Code Directly Eliminates Deploy Pipeline
I work with Claude Code in production mostly, then sometimes I push to save for the day (as a backup kinda?) But since it's in production I don't need the deploy pipeline anymore (laptop -> push -> GitHub -> webhook...
AI Layoffs Create Self‑inflicted Tech Debt Loop
“The causal chain is: I fired the engineers, the AI replaced the engineers, the AI broke what the engineers used to protect, and now the engineers I didn't fire must protect the system from the AI that replaced the engineers...
Hunter Alpha on OpenRouter: Just Okay So Far
So far, Hunter Alpha, the new mystery model on OpenRouter is only okay. Some examples of the Lem Test and the Sparks TiKZ unicorn. https://t.co/iXLx1jnoRe
AI Agents Auto‑review Code Before Humans See Diff
I type /commit and my code reviews itself - three agents fix issues before two more even see the diff. One markdown file, twelve repositories. https://t.co/rbbO2Cy10R
AI Co‑pilot: Adopt or Risk Obsolescence for Radiologists
AI is rapidly improving at reading medical images. But the real question is not whether it will replace radiologists. Most experts see AI becoming a powerful co-pilot: spotting patterns faster, flagging anomalies earlier and helping radiologists focus on complex diagnoses and...
Databricks Buys Quotient AI to Boost Genie Tools
Databricks acquires startup Quotient AI, a software to diagnose and evaluate AI agents. Databricks plans to utilize the acquisition to improve its Genie Code and Agent Bricks tools. Quotient AI was founded in 2023 by Julie Neagu and Freddie Vargus. The...
Built AI-Powered Options Screener in 35 Minutes
holy shit, @perplexity_ai computer build an options screener with AI built in for my parents in 35 minutes
AI Misplaces Funnel Stages, Requires Manual Correction
Gamma 5, Dave 0 in the "funnel wars" where I have tried 15 different prompts to get it to point the marketing funnel downwards. Let's not forget that it actually re-ordered the layers as well, somehow putting conversion at...
AI-Driven CX Insights Power Business Transformation
.@IvanK_NH has published a new @NHInsight blog post – 'Movate’s BizGenie: AI-powered CX Insights as Transformation Lever' https://t.co/NfXBnGkc4W #CXservices #CX #AI #GenAI @Movate_Inc https://t.co/fs3N0jTrkw
Humans Will Continue Overseeing AI for Years
This is a pretty good way of thinking about this, and leads to the conclusion that humans will remain as overseers of AI for quite a long time to come.
AI Lowers OSINT Barriers, Turns Iran Conflict Into Theater
How AI is turning Iran conflict into theater There’s a confluence of factors at play. AI coding tools mean people don’t need much technical skill to assemble open-source intelligence anymore, and chatbots can offer fast, if dubious, analysis of it. Yup.....
Bubeck's 2023 AGI Sparks Paper Still Holds Up
The 2023 “Sparks of Artificial General Intelligence” paper by Sébastien Bubeck @SebastienBubeck is aging very well.
Master Agentic Engineering and Gemini CLI's Plan Mode
Seroter Daily Reading List – March 11, 2026 (#739): Today’s links look at the 8 levels of agentic engineering, using “plan mode” in the Gemini CLI, and how to quash your fear of messing up. https://t.co/3J5m1ku9c4
AI Firms Plan $4.1T Capex, Outspending US Non‑financials
The 2026 OECD Debt Report shows the 9 major AI players have forecasted capital expenditures of $4.1 TRILLION for 2026-2030. That’s $1.1 TRILLION MORE than total capital expenditures by all US NON-FINANCIAL companies in 2025. WELCOME TO THE AI BUBBLE. https://t.co/YzxlQOpkiV
Future AI Will Constantly Rewrite Perfect Software for Users
One of the things everyone believes about software is that "once it works you shouldn't touch it." But what if that is wrong? What if in the world of infinite software that is very very good and smart and well tested...
AI Danger Lies in Centralization, Not Sentience
You're worried about the wrong AI problem. While everyone debates whether AI gets sentient and wipes us out, a handful of companies have already become the most powerful entities in the world. All the training data, all the compute, all the...
AI Builds My Workshop Landing Page in Real-Time
Current Situation: Drinking coffee on my patio watching Claude build out a landing page for my new workshop ‘Think with AI’ in Webflow. This is what work looks like now 👌 https://t.co/as5kMj9PPB
AI Use Risks NDA Violations and Data Exposure
Perplexity has changed my way of work. Love these guys. But not sure we use the word "secure" the same way. Strong protections on data, but both Perplexity and other models they exchange data with open up new threat...
Future Software: Managing AI Agents, Not Human Interfaces
Software will look very different in the future for most knowledge work. Our software was designed for people to do most of the work, but now agents will be doing a large portion of those tasks and don’t need any...
Human Social Platforms Translate Directly to AI Agents
Any major platform that worked for humans will probably work for agents. Agent to agent (A2A). Facebook/Reddit for agents (Moltbook) Ebay for agents (already a few of these) SEO for agents (Mintlify) Review sites for agents The list goes on and on.
AI Cuts Marginal Tech Hires Amid Fluctuating Job Market
AI eliminates the marginal hire. Tech job openings are down 45% from the 2022 peak, but up 16% since the start of 2026 - from 227k to 264k. Why the narrative violation? https://t.co/3QUv8XdkD4
AI Copilots Fail without User Training and Mindset Shift
Why are tools like Microsoft Copilot struggling for traction? Because they demand a change in user behavior without the necessary training. AI isn't intuitive like a Google search—it requires a new way of thinking about work. 🧠 Full analysis: https://t.co/EDcbE2J13x #CIO #AI...
AI Cuts Marginal Hires, Flattening Tech Workforce
AI eliminates the marginal hire. Tech job openings are down 45% from the 2022 peak, but up 16% since the start of 2026 - from 227k to 264k. Why the narrative violation? Companies are hiring again, just fewer people than...
Enterprise AI Success Requires Coordinated Usage
Enterprises are going to have to solve the problem of coordinating AI usage to achieve any real business benefits. Good read by @gsivulka . Also - big part of why I joined the BOD of @Larridin w/ @rfradin @jlarrison who are...
Coding Becomes Conversation, Not Skill, With Replit Agent 4
Replit Agent 4 just proved coding is no longer a skill. It's a creative conversation. The developers still arguing about frameworks are already behind.
AI Frees Us to Collaborate More on Whiteboards
AI Agents are handling the busywork, and something I've been saying is happening: humans are spending MORE time whiteboarding together, not less.
AI Can't Beat Markets; Hedge Funds Already Arbitrage Any Edge
While I think there are some great use cases for AI to do investment research (got an investment thesis, let it help crunch the data, or sift through a zillion 10-K filings to spot which companies match your criteria), I...
Agents Need User Interaction History, Not Just Code
Banger of a piece, from start to finish. 🔥 "We have handed agents the codebase, the tests, the docs, the specs, the commit history." "We haven't handed them the one artifact that actually survived — the record of what users have been...
MCP Servers Add Governance, Auditability, and Reuse
The real value of MCP servers isn't just connecting agents to tools. It's providing a governed layer for prompts, tools, and data access: curated capabilities, auditable actions, and reusable patterns teams can trust. #AI #CIO https://t.co/7dcoLIKa0K
Perplexity API Becomes Full‑stack, Model‑agnostic Search Platform
i've been saying this to whoever listens but the Perplexity API is an insanely great product. Sonar is probably the best web search API. And now it has a bunch more functionality. Excited to play around w/ it
LuxTTS Clones Any Voice Locally in Seconds
You can now clone any voice on a 4GB GPU. LuxTTS just killed the "you need ElevenLabs" excuse. It clones voices from 3 seconds of audio at 150x realtime speed. Fits in 1GB VRAM. Faster than realtime even on CPU. → 48khz output vs...
Perplexity API Powers Free Autonomous Agents via Telegram
.@perplexity_ai users can use their API keys right now with @AnimocaMinds Agents just spin up your agent (free) & tell them to load "Intelligence_Research_v1" (D5C1B743-1F13-F111-AD1D-0EA9A5017E89) & you have a Perplexity powered autonomous Agent operating via email @telegram +
Apple Lands AI Infrastructure Win with OpenClaw
$AAPL wins bigly here. Between this and OpenClaw it just found its AI infra play. 🙃
Self‑driven AI Adoption Spreads Companywide in Days
We had an All Hands at @eightsleep focused on AI. Less than a week later, the adoption was immediate. The progress was real, and most of it happened without anyone asking for help, the teams didn't wait for guidance. They figured...
AI Revives Dead Brands with Personal, Emotional Touch
Sears catalogs were once the engagement standard. We use AI to revive @DeadBrands with a personal touch. We transform data into emotional connections. Tell us what you want to hear from us. https://t.co/ZsEwnG4p2n https://t.co/AgJNZPQ38m
Claude Now Syncs Excel and PowerPoint Across Files
Wow. 🚨 Claude for Excel and Claude for PowerPoint now sync seamlessly. Have multiple files? Claude shares full context across them. It: → pulls sheet data → updates decks instantly → supports Bedrock/Vertex It even saves workflows as 1-click team 'Skills' 🤯 link 🧵↓ https://t.co/10E2iqSO9z
LLMs Skew Toward Western Morals, Ignoring Global Values
LLMs overemphasize moral concerns common in Western societies and underestimate values more prominent elsewhere. These distortions likely stem from cultural biases in training data and carry societal implications and risks https://t.co/0nacfBcoA7 https://t.co/FHZCLnu1Kn
AI Will Augment, Not Delete, SaaS for ROI
Stop looking for the "Delete" button on your SaaS contracts. The future isn't AI replacing SaaS—it’s AI augmenting it. Those who understand the difference will win the ROI game. 🏆 Get the full take: https://t.co/Se537FlJwE #CIO #Cloud #AI #EnSW
Netflix May Spend $600M on Affleck's AI Studio
Netflix Could Pay as Much as $600 Million for Ben Affleck’s AI Film Start-Up: Report https://t.co/EMp2y3V3u0 via @variety
AI Needs Human Oversight, Not Blame for Outages
If you are distressed because AI is causing outages at AWS well… don’t jump to conclusions like everyone did with the whole slew of S3 bucket debacles. Yes there will be problems as people learn how to use this new...
Fortune 500 Must Shift AI Focus to Epistemology
The Fortune 500 thinks about AI through the lens of efficiency and effectiveness: What they should be thinking about is epistemology.
LLMs Turn Writing Into Collaborative Reasoning Journey
Largely agree with the premise of this, in particular that often the best writing appears to be coming from someone who is trying to figure something out and takes you on that trip with them (The Flannery O'Connor quote is...
Mac Mini Hosting for Agentic AI Benefits MacStadium
Didn't have Mac Mini's as being the host computer for agentic AI on my bingo card. Looks like they are hosting a Mac Mini for you as part of the service. Maybe a good play for @MacStadium ?
AI Video Demand Lags as OpenAI Embeds Sora in ChatGPT
#OpenAI is putting #Sora inside #ChatGPT. Existing data suggests consumer demand for AI video may underwhelm relative to ChatGPT's 920 million weekly chatbot users. For Disney, its licensed 200-plus characters will be available to these audiences under strict conditions that...
Bring Your Own Agents—The New Unsettling Tech Mantra
The most interesting (and scariest?) phrase I’ve heard so far this year……. “Bring your own agents”
Local LLMs Will Make Hardware Specs Critical by 2028
The top .1% of users are playing with local LLMS This will 10x every 12 months.... Until Apple, Dell and MSFT have one on your local device BY DEFAULT in 2028 This takes the local hardware spec race from irrelevant for 99%...
Gemini CLI's Plan Mode Safely Plans Changes Read‑Only
"With plan mode, @geminicli focuses first on analyzing your request, planning complex changes, understanding the code base or dependencies, and all this happens in a read-only mode safe from accidental changes or executions." https://t.co/TMVjSdxqOC https://t.co/HrGAbtrJ7J
AI Multi‑Agents Let One Person Do Five Jobs
I love ai era. You can work of 5 people with multi agents without cost of communication and resources
