Insights for AI builders
Tutorials, product updates, and ideas to help you build and ship AI applications faster.
Subscribe via RSS
Liquid AI's LFM3 Family: Text, Vision, Audio, Omni and Decision Models
Liquid AI shipped a full LFM3 lineup in one day, covering text, vision, audio, omni and decision models. Here's what's actually in the release.

LiquidAI LFM3-1.7B: A Local Agent Model That Fits on a Phone
LFM3-1.7B is a 1.73B-parameter edge model running in ~1GB at 4-bit, built for on-device agents, tool calling, and fine-tuning.

LiquidAI LFM3-1.7B: A 1GB Local AI Model Built for Agents
LFM3-1.7B is LiquidAI's tri-hybrid model that runs in about 1GB at 4-bit. Here's how it works and how to run it locally.

Is OpenBot Free? Pricing, Hosted Servers, and License Explained
OpenBot's desktop app is free under PolyForm Noncommercial. Here's what it costs to run, what hosted servers add, and who needs a commercial license.

How to Run Carbon-X Locally for DNA Variant Analysis
Carbon-X is a 14.8B open DNA language model that fits on one 80GB GPU. Here's how to set it up with vLLM or Transformers.

Running Local LLMs on 8GB RAM Laptops: What's Actually Possible
What quantized Qwen 3 models can realistically do on 8GB-RAM laptops with llama.cpp, based on real token speed and memory tests.

How to Sell Claude-Based AI Automation Systems to Businesses
Four real Claude-powered automation systems sold for $2,000 to $15,000, with pricing logic and pitching strategies for AI builders.

How to Run Baidu's Unlimited OCR 2 Locally for Document Parsing
Baidu's Unlimited OCR 2 parses long documents locally with SR-SWA attention. Here's how installation, VRAM use, and real tests actually hold up.

Unlimited OCR 2 Benchmarks: Where Baidu's Model Wins and Fails
Unlimited OCR 2 benchmark results on MP-DocBench and Marathon-DocBench, tested on handwriting, mixed languages, old prints, and 100-page documents.

Anthropic's S-1: What the IPO Filing Reveals About Its Business
Anthropic's S-1 filing shows 900% net revenue retention and a self-reinforcing AI strategy, alongside the safety contradictions it raises.

Claude Opus 5.5 vs GPT-6.1 Soul and Terra: 3D Model Test Results
Opus 5.5, GPT-6.1 Soul, and Terra were given the same prompt to build a 3D diecast RX-7. One result stood out from the rest.

AI Just Disproved the Exponential Time Hypothesis. Here's Why It Matters
OpenAI says its AI disproved a 27-year-old computer science belief, the exponential time hypothesis, in 10 days. Here's what that means.

GPT-6.1 Terra vs Soul vs Luna: Which One Should You Use?
OpenAI's GPT-6.1 Terra sits between Soul and Luna on price and speed. Here's a breakdown of when each model actually makes sense.

GrokBot vs OpenBot: Cloud or Local AI Teammates?
GrokBot runs persistent AI agents in xAI's cloud. OpenBot runs them on your own machine with your own model plans. Here's how they compare.

MOSS Image: The Apache 2.0 Text-to-Image Model With JSON Layout Control
MOSS Image is a 5B Apache 2.0 text-to-image model with strong text rendering and JSON layout control. Here's how to run it and its limits.

MOSS Image vs Nano Banana and Qwen Image: Text Rendering Compared
MOSS Image, a 5B open-weight model, scores near Nano Banana and GPT Image 2 on text rendering. Here's how it stacks up against rivals.

Nanbeige 4.3 33B: Benchmark Charts vs. a Real Debugging Test
Nanbeige 4.3 33B tops its own benchmark charts and claims Claude-level agentic skill, but a hands-on coding test tells a different story.

Nanbeige4.3-33B Benchmarks: How It Beats Qwen3.8-27B and Gemma4-31B
Nanbeige4.3-33B tops Terminal-Bench, DeepSWE and JobBench among ~30B open models, beating Qwen3.8-27B and Gemma4-31B on most agentic tasks.

OpenAI's Three AI Science Claims: What Was Actually Proven?
OpenAI says its AI disproved the exponential time hypothesis, solved a 1991 math conjecture, and advanced Yang-Mills theory. Here's what that means.

How to Run OpenBot Locally for Free AI Agent Teammates
A practical guide to installing OpenBot, connecting your existing ChatGPT, Claude, Gemini or Grok plans, and running persistent local AI agents.

How Shopify's Design Team Replaced Figma-First Work With AI Agents
Shopify's product design studio now prototypes with Codex and Claude Code instead of Figma. Here's how that workflow actually works.

What Does It Mean to Be a 'Manager of Intelligence' at Work?
AI is becoming a third economic input alongside labor and capital. Here's what it means to manage it as an individual employee.

Could AI Voices Like Dessa Make Phone Scams Worse?
Dessa's full-duplex AI voice sounds fully human. Here's why that raises the risk of phone scams and why detection tools struggle to keep up.

Carbon-X Benchmarks: How It Stacks Up Against GPT-6.1 and Opus
Carbon-X hits 66.7% average across five DNA benchmarks, nearly matching GPT-6.1 Sol at roughly 1/1,800 the inference cost.