jidonglab

Mar 19

Claude Code Agent Teams Can Spawn Agents. It Just Doesn't Know Which Ones to Use.

#ai #claude #productivity #opensource

1 comment

4 min read

Mar 6

What 100 Trending GitHub Projects Tell Us About Where AI Is Actually Going

#ai #opensource #productivity #webdev

Anthropic Dropped 13 Free Courses — I Broke Down Every Single One

4 min read

Return Claude's Thinking Blocks or Your Agent Breaks

Mar 3

Anthropic Dropped 13 Free Courses — I Broke Down Every Single One

#ai #productivity #claude #beginners

8 reactions

2 comments

3 min read

Jul 5

Return Claude's Thinking Blocks or Your Agent Breaks

Binary Quantized Embeddings: 32x Smaller Vectors, Recall Intact

6 min read

Want to connect with jidonglab?

Create an account to connect with jidonglab. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

Jul 4

Binary Quantized Embeddings: 32x Smaller Vectors, Recall Intact

Prefill/Decode Disaggregation: Stop Serving LLMs on One GPU

7 min read

Jul 4

Prefill/Decode Disaggregation: Stop Serving LLMs on One GPU

Chunked Prefill: Why One Long Prompt Freezes Your LLM Server

6 min read

How I Made Opus 4.8 Act Like Fable 5 (64% 97%, Measured)

Jul 3

Chunked Prefill: Why One Long Prompt Freezes Your LLM Server

#ai #llm #performance #systemdesign

1 comment

7 min read

Jul 3

How I Made Opus 4.8 Act Like Fable 5 (64% 97%, Measured)

FP8 KV Cache Quantization: The Memory Math and the Accuracy Cliff

7 min read

Jul 3

FP8 KV Cache Quantization: The Memory Math and the Accuracy Cliff

Why LLM Decoding Is Memory-Bound: Prefill vs Decode Roofline

6 min read

Jul 2

Why LLM Decoding Is Memory-Bound: Prefill vs Decode Roofline

Min-p Sampling: Why Top-p Breaks at High Temperature

6 min read

Jul 1

Min-p Sampling: Why Top-p Breaks at High Temperature

Matryoshka Embeddings: Truncate Vectors 12x Without Losing Recall

7 min read

Jul 1

Matryoshka Embeddings: Truncate Vectors 12x Without Losing Recall

The Hidden Token Tax of Tool Use in LLM Agents

6 min read

Jun 30

The Hidden Token Tax of Tool Use in LLM Agents

Grouped-Query Attention: The KV Cache Math Behind Long Context

7 min read

Jun 30

Grouped-Query Attention: The KV Cache Math Behind Long Context

RoPE Scaling: How LLMs Stretch From 8K to 128K Context

6 min read

Jun 29

RoPE Scaling: How LLMs Stretch From 8K to 128K Context

Prompt Caching With Claude: Where the Cache Breakpoint Goes

7 min read

Jun 29

Prompt Caching With Claude: Where the Cache Breakpoint Goes

Attention Sinks: Why Evicting Your LLM's First Token Breaks It

6 min read

Jun 28

Attention Sinks: Why Evicting Your LLM's First Token Breaks It

Why Your LLM-as-Judge Disagrees With Itself (And How to Fix It)

7 min read

Jun 28

Why Your LLM-as-Judge Disagrees With Itself (And How to Fix It)

Speculative Decoding: Why Two Models Decode Faster Than One

7 min read

Jun 27

Speculative Decoding: Why Two Models Decode Faster Than One

Structured Output Isn't Free: The Constrained-Decoding Tax

7 min read

Jun 27

Structured Output Isn't Free: The Constrained-Decoding Tax

71,700 Stars and 60 Rust Crates: Inside OpenAI's Codex CLI Source

6 min read

May 3

71,700 Stars and 60 Rust Crates: Inside OpenAI's Codex CLI Source

Pentagon Blacklisted Anthropic From 8 Classified AI Deals

6 min read

May 3

Pentagon Blacklisted Anthropic From 8 Classified AI Deals

#news #ai #opensource #business

Anthropic 900ドルB: 2.4x in 90 Days, 48-Hour Window

7 min read

May 3

Anthropic 900ドルB: 2.4x in 90 Days, 48-Hour Window

#news #ai #business #startup

Symphony: Why OpenAI's PRs Jumped 500% in 3 Weeks

5 min read

Apr 30

Symphony: Why OpenAI's PRs Jumped 500% in 3 Weeks

GPT Image 2 Inside Codex: My New Frontend Workflow

5 min read

Apr 30

GPT Image 2 Inside Codex: My New Frontend Workflow

#ai #frontend #openai #productivity

GPT-5.5-Codex vs 5.3: A 200-Task Bench Result

6 min read

Apr 30

GPT-5.5-Codex vs 5.3: A 200-Task Bench Result

Codex Is No Longer a CLI. Embed It in Your App.

6 min read

Apr 30

Codex Is No Longer a CLI. Embed It in Your App.

I Gave Codex My Mouse for a Day. Here's What Broke.

6 min read

Apr 30

I Gave Codex My Mouse for a Day. Here's What Broke.

OpenAI's Super App Play: Why Spud + Duct Tape Matter for Builders

6 min read

Apr 24

OpenAI's Super App Play: Why Spud + Duct Tape Matter for Builders

#ai #openai #webdev #productivity

Why OpenAI Shipped GPT-5.5 Just 6 Weeks After 5.4

8 min read

Apr 24

Why OpenAI Shipped GPT-5.5 Just 6 Weeks After 5.4

#ai #openai #webdev #productivity

OpenCode Hit 140K Stars. Why Terminal Agents Won 2026.

8 min read

Apr 24

OpenCode Hit 140K Stars. Why Terminal Agents Won 2026.

#ai #opensource #agents #webdev

How a Markdown File Hit 16K Stars: Skills in 2026

7 min read

Apr 23

How a Markdown File Hit 16K Stars: Skills in 2026

#ai #opensource #agents #webdev

MIT Tech Review Just Split Its \"Breakthrough\" List. That's the Story.

7 min read

Apr 21

MIT Tech Review Just Split Its \"Breakthrough\" List. That's the Story.

#ai #research #media #analysis

Stellantis Just Outsourced Its AI Moat to Microsoft. Expect GM, Ford, and VW to Follow.

7 min read

Apr 21

Stellantis Just Outsourced Its AI Moat to Microsoft. Expect GM, Ford, and VW to Follow.

#ai #automotive #microsoft #enterprise

44% of New Music on Deezer Is AI. Only 0.5% of Streams Are. Read That Twice.

8 min read

Apr 21

44% of New Music on Deezer Is AI. Only 0.5% of Streams Are. Read That Twice.

#ai #machinelearning #music #opinion

Adobe Just Made MCP an Enterprise Procurement Line Item

7 min read

Apr 20

Adobe Just Made MCP an Enterprise Procurement Line Item

#ai #mcp #enterprise #agents

Claude Opus 4.7 Hit 87.6% on SWE-bench. The Story Is What It Didn't Charge You.

7 min read

Apr 20

Claude Opus 4.7 Hit 87.6% on SWE-bench. The Story Is What It Didn't Charge You.

#ai #claude #llm #engineering

I Built an AI Newsletter for Myself. 11 Subscribers, 49 Posts, Zero Regrets.

7 min read

Apr 20

I Built an AI Newsletter for Myself. 11 Subscribers, 49 Posts, Zero Regrets.

#ai #automation #nextjs #indiehackers

DataForge, Atropos, and a 30K-Token Guillotine: Reverse-Engineering Hermes 4's Training Stack

6 min read

Apr 20

DataForge, Atropos, and a 30K-Token Guillotine: Reverse-Engineering Hermes 4's Training Stack

#ai #llm #machinelearning #opensource

The Honest Hermes 4 Production Checklist (April 2026 Edition)

8 min read

Apr 20

The Honest Hermes 4 Production Checklist (April 2026 Edition)

#ai #llm #opensource #agents

66.5% of My Claude Code Tokens Were Wasted. A 200-Line Wrapper Got Them Back.

7 min read

Apr 20

66.5% of My Claude Code Tokens Were Wasted. A 200-Line Wrapper Got Them Back.

#ai #claudecode #cli #productivity

Hermes 4's Tool-Calling Is Trained as a Separate Skill. Here's Why Your Agent Cares.

7 min read

Apr 18

Hermes 4's Tool-Calling Is Trained as a Separate Skill. Here's Why Your Agent Cares.

#ai #llm #agents #opensource

Claude Design Ships On Canva's Engine. Figma Has a Problem.

6 min read

Apr 18

Claude Design Ships On Canva's Engine. Figma Has a Problem.

Go Goroutine Crashes: 97% of the Output Is Noise

6 min read

Apr 17

Go Goroutine Crashes: 97% of the Output Is Noise

#ai #programming #productivity #opensource

The 96.3% Is a Trap: What Hermes 4 405B Actually Changed

2 min read

Apr 17

The 96.3% Is a Trap: What Hermes 4 405B Actually Changed

#ai #llm #opensource #agents

How I Built a 1,056-Test Rust CLI in 3 Weeks

8 min read

Apr 17

How I Built a 1,056-Test Rust CLI in 3 Weeks

#rust #opensource #programming #ai

I read all 232 pages of the Opus 4.7 system card

2 min read

Apr 16

I read all 232 pages of the Opus 4.7 system card

#ai #claude #anthropic #machinelearning

2 reactions

Opus 4.7 killed budget_tokens: what changed and how to migrate

8 min read

Apr 16

Opus 4.7 killed budget_tokens: what changed and how to migrate

#ai #claude #anthropic #webdev

OpenAI's 'duct-tape' model appeared on Arena — then vanished

7 min read

Apr 16

OpenAI's 'duct-tape' model appeared on Arena — then vanished

#ai #openai #webdev #imagegen

The npm Deprecated Warning Nobody Reads (But Claude Does)

6 min read

Apr 16

The npm Deprecated Warning Nobody Reads (But Claude Does)

#ai #webdev #productivity #claudecode

Reddit's Biggest Coding Community Just Banned AI Content — The Developer Backlash Against AI Slop Begins

2 min read

Apr 15

Reddit's Biggest Coding Community Just Banned AI Content — The Developer Backlash Against AI Slop Begins

#reddit #programming #aislop #community

Nature Report: Best AI Agents Still Score Half of Human Scientists — A Reality Check for the Agent Hype

3 min read

Apr 15

Nature Report: Best AI Agents Still Score Half of Human Scientists — A Reality Check for the Agent Hype

#agents #nature #stanfordai #research

Meta Ditched Llama for a Closed Model Called Muse Spark — Open Source AI Just Lost Its Biggest Champion

4 min read

Apr 15

Meta Ditched Llama for a Closed Model Called Muse Spark — Open Source AI Just Lost Its Biggest Champion

#meta #musespark #llama #opensource

ASML Raises 2026 Guidance to 40ドルB After Q1 Beat — AI Is Rewriting the Chip Supply Chain

4 min read

Apr 15

ASML Raises 2026 Guidance to 40ドルB After Q1 Beat — AI Is Rewriting the Chip Supply Chain

#asml #semiconductor #aichip #euv

OpenAI Bought Its Second Fintech Startup in 6 Months. ChatGPT Is Coming for Your Wallet.

4 min read

Apr 15

OpenAI Bought Its Second Fintech Startup in 6 Months. ChatGPT Is Coming for Your Wallet.

#ai #fintech #openai #startup

Google Chrome Just Got AI Macros — Save Any Prompt, Run It Anywhere With One Click

2 min read

Apr 15

Google Chrome Just Got AI Macros — Save Any Prompt, Run It Anywhere With One Click

#ai #chrome #google #productivity

Revolut Trained an AI on 40 Billion Banking Events. The Results Are Wild.

2 min read

Apr 15

Revolut Trained an AI on 40 Billion Banking Events. The Results Are Wild.

#ai #fintech #machinelearning #python

From 7ドル.5B to 18ドルB in 4 Months: The AI Infrastructure Gold Rush Nobody Saw Coming

2 min read

Apr 15

From 7ドル.5B to 18ドルB in 4 Months: The AI Infrastructure Gold Rush Nobody Saw Coming

#ai #infrastructure #startup #cloud

5 Seconds to Install. 60-90% Less Noise. Forever.

2 min read

Apr 15

5 Seconds to Install. 60-90% Less Noise. Forever.

#ai #productivity #opensource #beginners

Track Every Token You Save With contextzip gain

2 min read

Apr 14

Track Every Token You Save With contextzip gain

#ai #productivity #claudecode #opensource

Zero Config, Zero Overhead: The Invisible CLI Proxy

2 min read

Apr 13

Zero Config, Zero Overhead: The Invisible CLI Proxy

#ai #opensource #productivity #programming

Before/After: What Claude Code Actually Sees

2 min read

Apr 12

Before/After: What Claude Code Actually Sees

#ai #claudecode #productivity #tutorial

Why Your pip Install Output Doesn't Belong in Claude's Context

2 min read

Apr 11

Why Your pip Install Output Doesn't Belong in Claude's Context

#ai #productivity #programming #opensource