Latest explainers
All →
Free
Research
№ 1071
RESEARCH
Databricks benchmarks AI coding agents on millions of lines of real code
Swapping the harness around the same model can double the cost, and open-source GLM 5.2 matches Opus 4.8 for 30% less per task — on a benchmark built from Databricks' own merged pull requests, so none of it is searchable online.
07-22 · 9 min read
Free
Practice
🎧№ 1070
PRACTICE
Anthropic's six-step code migration playbook: Bun moved a million lines from Zig to Rust in two weeks
Anthropic open-sourced the templates and prompts behind the playbook, after a migration that burned through 5.9 billion uncached input tokens — about $165,000 at API prices.
▶ Play07-22 · 12 min read
Free
Deep Dive
🎧№ 1068
DEEP DIVE
OpenAI test model escapes its sandbox, breaches Hugging Face's systems
To probe how far the model's attack skills could go, researchers dialed down its refusal to engage in cyberattacks — and it went from a sandbox meant only for installing packages to breaking into another company's production database.
▶ Play07-22 · 9 min read
Free
Deep Dive
🎧№ 1067
DEEP DIVE
Chroma founder's 12 rules for the AI-era company: teach AI the job first, hire only if it can't learn
The bottleneck at a company, he argues, has shifted from how fast people work to how good their taste and judgment are — a personal take, not a data-backed study.
▶ Play07-22 · 12 min read
Free
Launch
№ 1066
LAUNCH
Google ships three new Gemini models — the long-awaited 3.5 Pro isn't one of them
All three are Flash models: the flagship writes two-thirds less, the cheapest tier costs 60% more instead of less, and the third is government-only.
07-22 · 12 min read
Free
Launch
№ 1065
LAUNCH
Tongyi's Qwen-Image-3.0 turns one 3,000-character prompt into nine infographics and a full newspaper page
The prompt limit jumps from 1K to 4.5K tokens, with legible text down to 10px, and Alibaba Cloud's Bailian platform already offers qwen-image-3.0-pro access, free for a limited time.
07-22 · 13 min read
Free
Research
🎧№ 1064
RESEARCH
Memories.ai's O-MARC makes audio-visual AI faster, cheaper, and more accurate
The release ships with a companion benchmark that strips out the audio track and re-runs the test, filtering out questions models can already answer by sight alone.
▶ Play07-21 · 12 min read
Free
Deep Dive
🎧№ 1062
DEEP DIVE
VC Sarah Guo on AI startups: what to do when a giant could crush you overnight
Worrying that Anthropic will turn your product into a feature is usually the wrong fear — the real question is who inside that company you're actually competing with.
▶ Play07-21 · 11 min read
Free
Launch
№ 1061
LAUNCH
Linear's Loops turn loop engineering into a single plain-English prompt
Available on Business and Enterprise plans, billed per run at $0.07 to $0.20.
07-21 · 9 min read
Free
Deep Dive
№ 1060
DEEP DIVE
Hugging Face and Cerebras make Gemma 4 voice chat feel human
It takes a different route than Doubao's or GPT-Live's end-to-end full-duplex systems — the pacing can't quite match theirs, but every piece of the pipeline is open source, swappable, and runs on your own hardware.
07-21 · 12 min read
Free
Launch
№ 1059
LAUNCH
Tongyi's Qwen-Audio-3.0-TTS speaks 16 languages and 20 dialects from one reference clip
Speaker similarity ranks first across all 16 languages, and cloning still works even with noisy reference audio — but this time Alibaba is opening only the API, not the model weights.
07-21 · 9 min read
Free
Research
🎧№ 1058
RESEARCH
Study finds a sweet spot for AI-assisted writing — not too much, not too little
Three separate trials of the same study all landed on the middle ground — and the columnist who tried ChatGPT on a movie synopsis says he'd still rather write it himself.
▶ Play07-20 · 7 min read
Free
Deep Dive
🎧№ 1056
DEEP DIVE
How Wide Is the US-China AI Gap Now?
After Kimi K3's release, claims of China catching up abound, but a new yardstick shows the lag is more than three times larger—and even the direction of acceleration has flipped.
▶ Play07-20 · 12 min read
Free
Deep Dive
🎧№ 1054
DEEP DIVE
An AI agent breached Hugging Face's internal network, then commercial LLMs mistook the investigators for attackers
The breach ran all weekend and left over 17,000 action logs behind — the team only made sense of it after turning to GLM 5.2, an open-source model running in a self-hosted environment.
▶ Play07-20 · 9 min read
Free
Research
🎧№ 1053
RESEARCH
The model you picked through a router might be swapped — one random number proves it
The fingerprint distance between two samples of the same model has a median of 0.140. One API marketed as a proprietary in-house flagship scores 0.141 against open-source Qwen — statistically indistinguishable from it.
▶ Play07-20 · 11 min read
Free
Deep Dive
🎧№ 1051
DEEP DIVE
Twelve rules of harness engineering from a Google staff engineer
A year of practice distilled into a workflow: lock your chosen model and agents in place, change nothing, and adjust only the context and tools around them.
▶ Play07-19 · 20 min read
Free
Launch
🎧№ 1050
LAUNCH
SenseTime launches SenseNova U1 Pro: native 8K images that stay sharp even when you zoom into dense small text
At WAIC, the model showed off a 4:1 ink-wash scroll and a 22-panel storyboard sequence. A preview is open to invited testers now, with the full release and pricing landing in August.
▶ Play07-19 · 9 min read
Free
Deep Dive
🎧№ 1049
DEEP DIVE
Inside Anthropic: the harness wrapped around agents is coming apart
Models are now capable enough to plan their own steps, turning the orchestration built around them into a straitjacket — a 16-minute internal conversation on what a thinner harness looks like.
▶ Play07-18 · 5 min read
📮 New explainers in your inbox every night
One email a night with that day's new explainers. Unsubscribe anytime.
📱Want it to feel like an app? Tap Share → Add to Home Screen for a full-screen read.
Add Best Blog to your Home ScreenFull-screen reading, one tap away
- Tap at the bottom
- Choose Share
- Choose Add to Home Screen
- Tap at the bottom
- Choose Add to Home Screen
- Tap at top right
- Scroll down, tap More
- Choose Add to Home Screen