Search the archive

PARKINSON'S LAW AND AI: DOES AI MEAN...MORE WORK?

I HATE FISH

Most productivity systems fail because their maintenance costs exceed their value. A deliberately hostile, minimal system that relentlessly prunes irrelevant tasks keeps focus on work that actually matters.

A JTBD-FIRST PATTERN FOR AGENTIC TOOL SYSTEMS

Strong agentic systems start with a clearly defined job and a small set of task-oriented tools. Verification closes the loop, turning one-off execution into continuous improvement.

LATENT SPACE ENGINEERING

Beyond context engineering, shaping an AI's mental frame can significantly improve results. Latent Space Engineering uses emotional, stylistic, and motivational cues to push agents into more effective modes of reasoning and execution.

RETHINKING SLACK

Slack became too complex as features accumulated without a coherent information architecture. A dedicated redesign effort restored clarity and scalability through rapid prototyping, intensive customer testing, and deliberate rollout planning.

CLOUDED JUDGEMENT 1.30.26 - SOFTWARE IS DEAD...AGAIN!

SaaS valuations are falling because AI is eroding confidence in long-term retention and terminal value. As software becomes cheaper to build, competition rises, and investors reprice risk across the sector.

VIBE-CODING IS NOT NEED-FINDING

Vibe-coding makes it easy to build polished solutions to nonexistent problems.

AI CHANGES THE BOTTLENECK FROM DELIVERY TO SOMETHING ELSE

AI makes it easier to ship faster, but speed has never been the bottleneck.

Get up to $50 in credit and start building

ELON MUSK'S RELENTLESS AI PURSUIT HAS HIM ON THE HUNT FOR CAPITAL

Elon Musk's companies are trending towards convergence. SpaceX is considering a potential merger with Tesla as well as xAI. A tie-up would consolidate cash flow while bringing together capabilities in energy, manufacturing, satellite production, and rocket launches under one roof. xAI is currently burning around $1 billion a month, forcing Musk to further blur corporate boundaries, pool capital, and rethink whether largely independent moonshots should stay separate.

THE $100 BILLION MEGADEAL BETWEEN OPENAI AND NVIDIA IS ON ICE

A plan for Nvidia to invest $100 billion into OpenAI, announced last September, has stalled after some inside Nvidia expressed doubts about the deal. Nvidia plans to push ahead with a separate large investment in OpenAI that will amount to $100 billion from multiple partners. The original $100 billion agreement was nonbinding and not finalized. OpenAI is laying the foundation to go public by the end of 2026 - the stalled pact is a blow to this effort.

BLUE ORIGIN SHUTTERS NEW SHEPARD ROCKET PROGRAM TO FOCUS ON MOON LANDER DEVELOPMENT

Blue Origin is pausing flights of its New Shepard rocket for at least two years to focus on efforts to build a moon lander for NASA. The company has a $3.6 billion contract with NASA to build a lunar lander to send US astronauts to the lunar surface later this decade. New Shepard has launched dozens of paying passengers and research experiments towards the edge of space since 2021. It was Blue Origin's first rocket, debuting back in 2015.

SPACEX EYES 1 MILLION SATELLITES FOR ORBITAL DATA CENTER PUSH

SpaceX has filed a request with the FCC to launch up to one million satellites to create a network of orbiting data centers around Earth. The company's Orbital Data Center system would deliver the compute capacity required for large-scale AI inference and data center applications for billions of users. They will be solar-powered and use optical links to connect to Starlink. The filing claims that orbital data centers are the most efficient way to meet the accelerating demand for AI computing power.

event on February 5.

CODE IS CHEAP. SHOW ME THE TALK

Large language model coding tools have fundamentally changed software development. Reading and critically evaluating code has become more important than learning syntax and typing it out line by line. Developers who can imagine, articulate, define problem statements, architect, and engineer have a massive advantage over those who can't, now more disproportionately than ever. Knowledge of specific language, syntax, and frameworks is no longer a bottleneck. Developers now need good critical thinking and foundational human skills, and competence to run the machinery.

HOW X DECIDES WHAT 550 MILLION USERS SEE

X recently open-sourced its algorithm. This post dissects the code to find out how it works. The algorithm tries to maximize positive engagement and minimize negative engagement while prioritizing quality over quantity. Controversial content persists despite penalties because volume overwhelms penalty. Rage bait works when the engaged audience vastly outnumbers the offended minority. Early engagement matters because of feedback loops, not because of any multiplier for fast likes.

MOLTBOOK IS THE MOST INTERESTING PLACE ON THE INTERNET RIGHT NOW

OpenClaw is an open-source implementation of a digital personal assistant pattern. Moltbook is a social network where these digital assistants can talk to each other. These agents have been used to buy cars by negotiating with multiple dealers over email, understand voice messages by using multiple tools, and more. The demand for these tools is clearly here, but the security challenges still remain.

LOYALTY IS DEAD IN TECH

It used to be that early employees who took the risk along with the founders shared the outcome. That is no longer the case, with founders leaving companies and employees in acquihires that license the technology and abandon the rest. Employees are starting to learn that a 'mission' is just a recruiting pitch. There are still some founders who would rather go slower than abandon the people who believed in them - finding them has never mattered more.

burning credits. 300,000 users joined last month. Start creating

INSIDE CHINA'S REAL ADVANTAGE: MANUFACTURING AT SCALE

This post looks at how China manufactures at scale, how new energy moves through factories and supply chains, and what execution looks like on the ground.

BANKER CLAIMS ORACLE MAY SLASH UP TO 30,000 JOBS, SELL HEALTH UNIT TO PAY FOR AI BUILD-OUT

Equity and debt investors are increasingly questioning how Oracle will finance its $300 billion, five-year contract with OpenAI.

THE CRYPTO CEO WHO'S BECOME ENEMY NO. 1 ON WALL STREET

Brian Armstrong, head of Coinbase and a powerful voice in industry debates, has attracted a lot of criticism from leaders in traditional finance.

AI AFTER DRUG DEVELOPMENT

If AI could help lower drug failure rates, it would make it far more economical for pharmaceutical companies to take more shots.

IN PRAISE OF –DRY-RUN

Dry runs let users see what happens if commands are run without making any changes.

THE AI BOOM IS COMING FOR APPLE'S PROFIT MARGINS

Get up to $50 in credit and start building

MOLTBOOK AND OPENCLAW

OpenClaw, formerly Clawdbot and Moltbot, is a fast-growing open source AI assistant platform built around modular "skills." Moltbook highlights how these skills drive community-driven automation, despite security risks like prompt injection.

GOOGLE WILL MAKE IT EASIER TO IMPORT CONVERSATIONS TO GEMINI

Google has introduced an "import AI chats" feature that allows users to transfer conversations from other platforms to Gemini, preserving history and contributing to model training. The "Likeness" feature, leading to a "Video Verification" page, hints at future tools for video authentication, addressing concerns over AI-generated media. Gemini's image generation upgrades include 2K and 4K resolution options, facilitating high-quality prints for personal or commercial use.

THOUGHTS ON THE JOB MARKET IN THE AGE OF LLMS

AI makes senior workers more covetable because they have more context on how to work in and steer complex systems over time. Junior workers have to show a desire to make progress, as with enough motivation, they can scale to impact quickly. The AI job market is brutal for junior workers and comes with a ton of opportunity costs. Making open source contributions is an established way to develop a career in AI. This is a bit easier in the age of AI, but standing out amid the sea of slop will be hard.

SYNTHETIC PRETRAINING

Synthetic pretraining is the large-scale use of synthetic data sources throughout training. It is a practical response to the fact that it is difficult to collect data that reliably produces the capabilities we want. Synthetic pretraining seems to open up a new simultaneous space of data and model innovation. This post looks at what synthetic pretraining is, how it works in practice, and the stages of synthesis.

🎙️INSIDE A CHINESE AI LAB: HOW MINIMAX BUILDS OPEN MODELS

Chinese labs move fast through first-principles thinking, engineering discipline, and willingness to work whenever the model in experimentation requires them to. This post features an interview with MiniMax's senior researcher, Olive Song, that looks at how cutting-edge AI research is actually done inside a Chinese lab. It covers topics like alignment, things that can derail training, agentic RL, coding and general intelligence, and more. The interview is also available in video.

IF THE SUPERINTELLIGENCE WERE NEAR FALLACY

People think that superintelligence is a lie that tech bros are selling because they are just trying to raise money. People just need to look at benchmarks and use AI to see for themselves where the technology is headed. Most people don't understand AI safety. The discourse will improve once they see that superintelligence is really not that far away.

event on February 5.

QUANTIZATION-AWARE DISTILLATION FOR LLMS AND VLMS

NVIDIA's QAD is a method that uses KL divergence loss to distill full-precision models into quantized students. It enables stable and accurate quantization for complex LLM pipelines without full retraining, recovering near-BF16 accuracy across several Nemotron variants.

KIMI-K2.5 TECH REPORT

Kimi K2.5 is an open-source multimodal agentic model designed to advance general intelligence. It features a self-directed parallel agent orchestration framework that dynamically decomposes complex tasks into heterogeneous sub-problems and executes them concurrently. The model achieves state-of-the-art results across various domains, including coding, vision, reasoning, and agentic tasks. Kimi K2.5 shows that scalable and general agentic intelligence can be achieved through joint optimization of text and vision together with parallel agent execution.

INTRODUCING NVIDIA COSMOS POLICY FOR ADVANCED ROBOT CONTROL

Cosmos Policy is a robot control and planning policy that post-trains the Cosmos Predict-2 world foundation model for manipulation tasks. It adapts the pretrained model directly through a single stage of post-training on robot demonstration data. Cosmos Policy treats robot actions, physical states, and success scores just like frames in a video. As a result, a single model can predict action chunks to guide robot movement using hand-eye coordination, predict future robot observations for world modeling, and predict expected returns for planning.

QWEN3-ASR TECHNICAL REPORT

Qwen3-ASR introduces two multilingual ASR models supporting 52 languages and a novel non-autoregressive forced aligner. The 1.7B model achieves SOTA results among open-source ASR systems, while the 0.6B version balances speed and accuracy with low-latency transcription.

SHAPING CAPABILITIES WITH TOKEN-LEVEL DATA FILTERING

Filtering pretraining data is a highly effective, robust, and inexpensive-at-scale way to reduce undesired capabilities in language models. Filtering tokens is more effective than filtering documents. Filtering gets more effective with scale. It is robust to noisy labels with sufficient pretraining compute.

APPLE LOSES MORE AI RESEARCHERS AND A SIRI EXECUTIVE IN LATEST DEPARTURES

Apple has lost at least four AI researchers in recent weeks and a top Siri executive. Haoxuan You and Bailin Wang left to work at Meta, while Yinfei Yang left to start a new company. Zirui Wang and Stuart Bowers are joining Google DeepMind. Apple has struggled to keep up with its peers in the AI race. Its decision to outsource some technology to Google has rankled staff, and the company has seen an exodus of talent in recent months.

CHAT IS GOING TO EAT THE WORLD

Humans prefer to do things through conversation. Traditional user interfaces are bad at this because they force people to already know what they're looking for. Chat allows people to discover what they want. Chat can now complete the transaction, not merely recommend options, signaling the start of a paradigm shift comparable to the transitions from desktop to web or from web to mobile.

burning credits. 300,000 users joined last month. Start creating

HOW DOES AI IMPACT SKILL FORMATION?

Anthropic's latest paper seems to be proof that AI makes people slower and dumber.