标签：#tools

The memory layer that decides what's worth remembering Discussion | Link

MCP Was a Mistake. Here Are 200,000 Tokens That Prove It. "mcp were a mistake. bash is better." — Peter Steinberger, OpenClaw founder I didn't want to believe it either. MCP was supposed to be the USB-C of AI — one protocol to connect everything. Anthropic, OpenAI, Google all backed it. 97 million monthly downloads. 17,000 servers. But then I measured what MCP actually does to your context window. The Setup I connected 10 popular MCP servers to a token counter. Here's what happened before I typed a single word: Server Tools Tokens Injected Filesystem 11 3,847 Brave Search 8 2,103 Sequential Thinking 3 890 Memory 9 2,567 Puppeteer 15 5,890 Postgres 19 8,231 Notion 24 13,780 GitHub 28 12,440 Slack 22 14,672 Google Drive 31 47,293 Total 170 111,713 111,713 tokens. Before your first message. That's not a typo. Connecting 10 MCP servers to Claude means over 100K tokens of JSON schemas get injected into your context window. You haven't asked a question yet. You haven't made a tool call. The schemas are just... sitting there. The Math That Made Me Angry At Claude 3.5 Sonnet pricing ($3/M input tokens): Every conversation starts with 111K tokens of overhead: $0.33 20 conversations per day: $6.67/day 22 working days per month: $147/month Annual cost of JSON schemas: $1,764 That's more than a Claude Pro subscription. You're paying $1,764/year to read JSON braces describing tools you might never use. But Wait — It Gets Worse The 111K is just the schema injection. When you actually call a tool, MCP wraps the result: { "content" : [ { "type" : "text" , "text" : "{ \" file \" : \" app.py \" , \" size \" : 1024}" } ] } The actual content is 38 characters. The wrapping is 47 characters. 55% of your result tokens are JSON overhead. With 20 tool calls per conversation: Schema injection: ~111K tokens Result wrapping: ~18K tokens Total overhead: ~130K tokens per conversation Your $0.54 conversation now has 130K tokens that serve zero purpose. What Garry Tan Was Right About When YC's CE

Claude Code Is Burning Your Token Budget. Here's the Receipt. I found $2,500/year of hidden token waste in my Claude Code setup. It was the MCP servers. The Discovery Last week I noticed my Claude Code conversations were dying at around message 15. Context window full. The model starts forgetting earlier instructions. Tool calls fail. The conversation degrades into hallucination. I assumed it was my fault — too many messages, too much context. So I started measuring. Here's what I found: Session start: Claude system prompt: ~8,000 tokens MCP schema injection: ~111,000 tokens User's first message: 50 tokens ────────────────────────────────────────── Total before any work: ~119,000 tokens Remaining context: ~81,000 tokens I was starting every conversation with 60% of my context already consumed. The culprit wasn't my prompts. It was the 10 MCP servers I had proudly configured in my claude_desktop_config.json . The Receipts I measured each server's schema injection using tiktoken: Server Why I Installed It Token Cost Times Used/Week GitHub PR reviews, issues 12,440 3 Slack Message reading 14,672 0 Google Drive Doc access 47,293 1 Notion Knowledge base 13,780 2 Postgres Query DB 8,231 4 Puppeteer Screenshots 5,890 0 Filesystem File access 3,847 15 Brave Search Web search 2,103 5 Memory Context persistence 2,567 0 Sequential Thinking Reasoning 890 2 Total 111,713 Look at the "Times Used/Week" column. Three servers were used zero times. Two more were used once or twice. But every single one of them was injecting 100% of its schema into every conversation. I was paying $0.33 per conversation — $2,500/year — to load schemas for tools I barely used. The Moment I Realized Everyone Has This Problem I posted my findings on Bluesky. Within hours: "I had the same issue. Removed 6 MCP servers and my conversations went from dying at message 15 to lasting 40+ messages." — @developer1 "GitHub MCP is 12K tokens but Claude Code already has gh CLI built in. Why did I install it?" — @dev

Garry Tan Was Right: "MCP Sucks Honestly." I Have the Token Receipts. "MCP sucks honestly. Context window eats too much, auth is a mess. I wrote a CLI wrapper in 30 minutes and it works better." When YC's CEO says this on X, people listen. But nobody had the data to back it up. Until now. What Garry Tan, Perplexity's CTO, and 97 Million Downloads Can't Hide Three things happened in the last 6 months that changed how I think about MCP: Peter Steinberger (OpenClaw founder) tweeted: "mcp were a mistake. bash is better." Eric Holmes wrote "MCP is dead. Long live the CLI" — it hit HN frontpage Denis Yarats (Perplexity CTO) publicly announced they're replacing MCP with REST API + CLI internally Garry Tan (YC CEO) replied: "MCP sucks honestly" The community split into two camps: "MCP is dead" — CLI is simpler, cheaper, faster "MCP is fine" — 97M downloads, 17K servers, it's the standard Both are wrong. The problem isn't MCP. The problem is what MCP does to your context window. The 47,000-Token Problem Nobody Measured I connected 10 MCP servers to a token counter. Here's what I found: MCP Server Tools Token Cost Equivalent Sequential Thinking 3 890 This blog post Brave Search 8 2,103 A short email Filesystem 11 3,847 A README Memory 9 2,567 A meeting note Puppeteer 15 5,890 A chapter of a book Postgres 19 8,231 A whitepaper GitHub 28 12,440 A court filing Notion 24 13,780 A legal contract Slack 22 14,672 A novella chapter Google Drive 31 47,293 Half of a novel Total 170 111,713 A short book One MCP server — Google Drive — injects 47,293 tokens into your context before you ask a single question. The entire works of Shakespeare is 900K tokens. Google Drive's schema is 5% of Shakespeare. For listing files. The Cost Breakdown (So You Can Get Angry Too) At Claude 3.5 Sonnet pricing ($3/M input tokens, $15/M output): Scenario Tokens Cost Annual Cost 1 server (minimal) 3,847 $0.01/conv $4.40/yr 3 servers (common) 14,528 $0.04/conv $19.40/yr 5 servers (typical) 33,061 $0.10/conv $4

I Benchmarked 10 MCP Servers — One of Them Burns 47K Tokens Just to Say Hello 10 popular MCP servers. 847 tools total. 312K tokens of JSON schemas. One server alone wastes more tokens than a full GPT-3 conversation. Here are the results. What I did I installed the 10 most popular MCP servers from the official registry. Connected each one to a token counter. Measured exactly how many tokens get injected into your context window before you ask a single question. The servers: # Server Tools Token Cost 1 Filesystem 11 3,847 2 GitHub 28 12,440 3 Postgres 19 8,231 4 Puppeteer 15 5,890 5 Brave Search 8 2,103 6 Memory 9 2,567 7 Sequential Thinking 3 890 8 Slack 22 14,672 9 Google Drive 31 47,293 10 Notion 24 13,780 Totals: 847 tools across 10 servers 111,713 tokens of JSON schemas 200,000+ tokens including server status messages, headers, and error schemas That's right — connecting 10 MCP servers to Claude means 200K tokens of overhead before your first message . The worst offender: Google Drive Google Drive's MCP server exposes 31 tools. Each tool has deeply nested schemas for file operations, permission management, sharing, and search. The full schema dump: { "name" : "drive.files.list" , "description" : "Lists files in the user's Google Drive with optional filtering" , "inputSchema" : { "type" : "object" , "properties" : { "q" : { "type" : "string" , "description" : "Query string for filtering files..." }, "corpora" : { "type" : "string" , "enum" : [ "user" , "domain" , "sharedDrive" , "allDrives" ]}, "includeItemsFromAllDrives" : { "type" : "boolean" }, "orderBy" : { "type" : "string" }, "pageSize" : { "type" : "integer" }, "pageToken" : { "type" : "string" }, "spaces" : { "type" : "array" , "items" : { "type" : "string" }}, "supportsAllDrives" : { "type" : "boolean" }, "fields" : { "type" : "string" } }, "required" : [] } } That's ONE tool. 31 of them. At ~1,525 tokens per tool average. 47,293 tokens. Just for Google Drive. For comparison, the entire works of Shakespea

Search and act across all your work apps Discussion | Link

Control your company's internal AI agents and tools Discussion | Link

skillcheck is a static analyzer for SKILL.md files, the format agents like Claude Code, Copilot, Codex, and Cursor use to load reusable skills. It validates frontmatter, scores description discoverability, checks file references, enforces token budgets, and flags cross-agent compatibility issues. No network calls, no LLM calls, no file mutations. Runs as a CLI, a GitHub Action, or a pre-commit hook. pip install skillcheck skillcheck skills/ Latest pass was hardening and accuracy, not features. Here's what changed and why. Description scores went up. Skills that were scoring low because the scorer was broken will now see a jump in scoring. Median across the reference corpus went from 75 to 90. --explain-score also now tells you which pattern hits or misses instead of just a number. The score exists to predict whether an agent will actually find and trigger your skill, so a scorer that under-credits good descriptions defeats the point. The fix was validated against real-world skills, and the separation held: filler still scores 28-65, well-written descriptions 85-100. Corrupt files now fail cleanly instead of crashing. Before, a bad history ledger or non-UTF-8 skillcheck.toml above the skill dumped a Python traceback. It's now a clear error naming the file and byte offset (exit code 2). Config discovery walks up the directory tree, so one bad file could break every scan under it. Now every untrusted read (ingest, history, config) goes through the same guard before parsing, so they all reject the same way. README has been corrected in regards to token estimates. Without tiktoken, expect roughly 20-30% over-estimation, so install the extra if you're near a budget limit. The offline heuristic feeds the budget checks and its accuracy had never actually been measured, just assumed. It's benchmarked against tiktoken across the full corpus now, and the documented numbers are the measured ones. pip install "skillcheck[tiktoken]" The rest of the pass is invisible on purpose: f

Turning desktop chaos into curated spaces. Discussion | Link

Talk plainly and create macOS native automations Discussion | Link

One mouse, keyboard and second screen for all your computers Discussion | Link

A shared workspace where people and AI agents get work done Discussion | Link

Turn your GTM skills into recurring revenue Discussion | Link

Drift score for any GitHub repo Discussion | Link

Waymo’s Gemini integration in Ojai is being introduced through a limited beta rollout , not as a feature available to every rider. The company’s official updates describe early access for a restricted group of riders, followed by gradual expansion, while Gemini in the Ojai cabin continues to receive interface and product enhancements. Waymo first outlined the rollout in May 2026, saying it would welcome initial riders in select cities, including San Francisco, Phoenix and Los Angeles. Those riders would receive free rides as Waymo collected feedback, with access expected to expand over time. Waymo’s May 2026 Ojai rollout announcement is the primary account of that initial plan. The company’s July 2026 update characterized Gemini in Waymo as a beta feature in the Ojai cabin. It also referred to a redesigned user interface and ongoing enhancements. Taken together, the two announcements establish a phased deployment: early rider feedback first, followed by continued iteration and broader access over time. What Waymo has announced about Gemini in Ojai The official information supports a narrower view of availability than a universal launch. In May, Waymo described a limited group of early riders in three named cities. By July, the company was still describing Gemini in the Ojai cabin as a beta. Rollout stage Waymo’s description What it indicates May 2026 First riders in San Francisco, Phoenix and Los Angeles, with free rides for a limited group while Waymo collected feedback. Initial access was restricted and feedback-led. July 2026 Gemini in Waymo was described as a beta feature in the Ojai cabin, with a redesigned UI and ongoing enhancements. The in-cabin experience remained in active development. Waymo did not provide a universal-access date in the supplied material. Its stated approach is gradual expansion, which means availability should be assessed by the company’s current rider communications rather than by assuming that a beta feature has reached all users. Why

When merging CSVs in the browser, handling mismatched columns and quoted cells changes everything. Here's how filetools does it. Last week we shipped CSV merge/split/transpose tools for filetools, and the most interesting challenge wasn't CSV parsing - it was handling real-world data without a server. Here's how we handle the hard cases. The Problem: CSV files in the wild are messy. Columns don't always match. A cell value contains a comma and that comma is quoted. Headers are sometimes case-sensitive, sometimes not. When you build on a server, you can run a fast library and stream the result. In the browser, you have to make your merge operation deterministic from first load. Our approach: Column matching: Users specify which columns to merge on (e.g., "id" or "email"). We do a case-insensitive first pass, then check for exact matches. If no match exists, we warn the user and ask them to pick from the detected headers. This upfront clarity saves merge errors later. Quoted cell handling: We follow RFC 4180 strictly - a quote inside a quoted field is escaped as a double quote. Most CSV parsers get this wrong when they're quick. We use the csv-parse library (MIT) vendored into the site, same way we do with PDF and ZIP libraries. Column order: The merge operation respects column order from the first file, then appends any new columns from subsequent files. This is deterministic and reproducible. Why this matters for a browser tool: Server-based CSV tools hide their assumptions - you upload, they merge, you download. If a merge fails, you get an error message and no insight into why. Client-side, the user can see the detected headers, approve or correct them, and re-try immediately. That transparency matters when you're dealing with data that represents real records or transactions. What shipped this week: We added merge, split (by row count or column value), transpose, and comparison tools. The same deterministic, transparent approach applies to each one. Next question

OpenAI has formally rolled out a more flexible way for organizations to buy and use Codex. Eligible ChatGPT Business and Enterprise workspaces can add Codex-only seats on a pay-as-you-go basis, with usage billed through API-style token consumption rather than a fixed fee for each of those seats. The change matters because it separates AI coding access from a conventional per-user subscription model. A team can provision Codex-only members for development work while managing consumption through credits, rather than treating every user as a full ChatGPT seat. OpenAI also reduced the annual ChatGPT Business seat price from $25 to $20 and introduced promotional credits intended to lower the cost of onboarding new Codex users. OpenAI details the model in its official flexible Codex pricing announcement for teams . The initial announcement was made on April 2, 2026. A June 24, 2026 update clarified that existing pay-as-you-go seats remain in place, while the availability of new Business pay-as-you-go seats may change. How OpenAI's Codex team pricing works The central change is the availability of Codex-only seats within Business and Enterprise workspaces. These seats have no fixed seat fee . Instead, their Codex consumption is charged using API-style token rates. OpenAI's approach gives organizations a way to make Codex available to users whose primary need is AI-assisted coding, without requiring the same seat model used for standard ChatGPT access. This is not a claim that coding costs are universally lower. A pay-as-you-go model makes expenditure more directly dependent on usage, so the financial effect will vary by team and workload. The documented advantage is billing flexibility: organizations can align Codex costs with token consumption and use credits across supported workspace features where their plan allows it. Item ChatGPT Business seat Codex-only seat Pricing approach $20 per seat annually, reduced from $25 Pay as you go through API-style token consumption Fi

The AI port manager for Mac. Never run lsof again. Discussion | Link

Cloudflare, Astro, AI agents, GitHub Actions, issue triage, agentic AI, software architecture, open source, developer tools, AI automation, automated testing, human in the loop, agent workflows, GitHub, software engineering, AI software development, bug triage, continuous integration, developer productivity, autonomous agents, AI coding, Cloudflare Workers, Flue, triagebot By Leela Kumili