AI: Claude Tag by Anthropic: chat agent in Slack

another innovation... post Touring Test tool... 

Introducing Claude Tag \ Anthropic

Claude Tag is a new way for teams to work with Claude.
We’re starting on Slack, which Claude can join as a team member. Grant Claude access to selected channels, and connect it to whichever tools, data—and even codebases—you choose. Then, anyone in the channel can tag @Claude in, and delegate tasks to it while they focus on other work. Claude builds context by remembering relevant information from the channels it’s in, and can plan out tasks to complete in the future.
We see Claude Tag as the beginning of an evolution of Claude Code: it makes the model even more proactive, and it works better with a full team.

Anthropic is coming for EVERYTHING - YouTube by MattB

This video explores the implications of Anthropic's new feature, Claude tag, which integrates Claude directly into Slack (0:00). While initially appearing as a convenient tool for team collaboration, the creator argues it represents a significant shift in how companies operate.

Key takeaways:

• Integration as Infrastructure: Claude tag operates in an "ambient mode," reading conversations, documents, and context across the organization to become a persistent, AI-driven teammate (0:52 - 1:41).
• The "Trojan Horse" Risk: The creator warns that this leads to "context lock-in," where companies effectively rent their own institutional knowledge and operations back from Anthropic (9:16 - 10:18).
• Uncapped Costs: Unlike human employees with salaries, AI agents have tokenized, potentially infinite costs, which could create a competitive disadvantage for companies unable to afford massive token usage (10:52 - 11:21).
• The Future of Software: The video suggests that as AI agents begin to operate software interfaces directly, the value of traditional SaaS user interfaces may diminish, potentially turning software companies into mere database providers for these agents (11:45 - 13:07).

The creator concludes that while this evolution toward AI-native companies is inevitable and productive, it necessitates a focus on open-source models and model competition to prevent a single entity from monopolizing knowledge work (13:34 - 14:51).

Tag Claude in, right where you already work - YouTube

Anthropic’s Claude Tag is learning your company, one Slack message at a time | TechCrunch

Anthropic and Salesforce Announce New Claude to Slack Integration | Salesforce Ben

Salesforce and Anthropic have launched Claude Tag, a new method of connecting the two technologies by utilizing Claude in Slack.

Once installed into your Slack workspace, users can now type @Claude in any channel to instantly loop in Claude to collaborate on workflows, requests, and ideation in real time. This is what Salesforce is calling an “ambient” colleague in action.

a step to "AI native company"

IBM Debuts World’s First Sub-1 Nanometer Chip Technology

... featuring a revolutionary transistor architecture at the 0.7 nm...

architecture: Natural Ventilation

Lesson 4 - Natural Ventilation - YouTube

This video explores passive ventilation, a sustainable architectural strategy that uses natural forces like wind and thermal buoyancy to keep buildings cool, fresh, and energy-efficient without relying on mechanical systems like fans or air conditioning (0:00 - 0:35).

Key Concepts of Passive Ventilation:

• Primary Forces:

  • Wind-driven ventilation: Outdoor wind enters the building, pushing fresh air in and stale air out (0:37 - 0:47).
  • Stack effect: Utilizes the principle that hot air rises to escape through upper openings, while cooler air is drawn in from below (0:47 - 0:54).

• Architectural Strategies:

  • Cross ventilation: Positioning windows on opposite or adjacent walls to facilitate straight-through airflow (0:55 - 1:07, 1:54 - 2:03).
  • Vertical airflow: Using skylights, roof vents, or high-level windows to release hot air, creating a stack effect (1:11 - 1:27, 2:53 - 3:03).
  • Internal circulation: Using features like transom windows, ventilated partitions, or Jolly screens to allow air to move between rooms (3:27 - 3:39).

• Design Considerations:

  • Window Selection: Casement windows catch more wind than sliding ones; louvered windows offer controlled airflow; corner windows capture breezes from multiple directions (2:14 - 2:32).
  • Shading: Essential to prevent hot air intake. Strategies include horizontal overhangs, vertical fins/louvers, pergolas, and strategically planted deciduous trees (3:39 - 4:17).
  • Climate Adaptation: Designs must be tailored to the specific environment—for example, hot/dry climates benefit from night purging and courtyards, while hot/humid areas prioritize cross ventilation (4:17 - 4:34).

AI biology workflows: Proto by Stanford

 A high-level programming language for generative biology with Proto | bioRxiv

Stanford University brings with Proto orchestration to AI biology workflows → Brian Hie’s team released Proto, an open framework for composing AI biology models across DNA, RNA, proteins and ligands into one pipeline.

A high-level programming language for generative biology with Proto PDF

link from humanintheloop.online/p/i-stopped-writing-prompts-now-i-write-loops

AI summary 

The paper introduces Proto, a high-level, open-source programming language designed for generative biology. While traditional biological engineering relies on mixing and matching literal sequence parts from nature (often through trial-and-error), Proto establishes modularity at a functional and semantic level. It leverages generative AI models to bridge the gap between high-level functional intent and low-level biological sequences.

---

The Four Primitives

Proto unifies diverse biological AI models by breaking down design tasks into four core abstractions:

• Sequences: Typed variables representing physical strings of DNA, RNA, proteins, or ligands.

• Constraints: Scoring functions (ranging from basic stats like GC content to complex networks like AlphaFold) that evaluate a sequence's desirability.

• Generators: Procedures that propose candidate sequences (such as language models or diffusion models).

• Optimizers: Iterative loops (like MCMC or gradient descent) that guide the generator toward minimizing constraint scores.

---

Key Achievements & Validation

The authors demonstrate Proto's flexibility across multiple modalities and scales:

• Recapitulating Past Campaigns: They successfully reproduced diverse literature designs in silico, including symmetric protein homo-oligomers, de novo protein monomers, multi-modal CRISPR-Cas systems, 20-kb chromatin accessibility tracks, and antibody CDR designs.

• Experimental Validation:

  • RNA Introns: Designed alternatively spliced introns tailored to specific human cell lines, validated experimentally in cell cultures.

  • Promoter-Repressor Pairs: Achieved leading experimental success rates for synthetic protein-DNA design.

• AI Agent Integration: By coupling Proto with general-purpose AI coding agents, they showed it can generate complex biological designs (like cancer-targeting therapies and multi-step pathways) directly from natural language instructions.

Ultimately, Proto aims to make generative biological design highly structured, scalable, and accessible to researchers across varying levels of computational expertise.

OpenAI custom "chip" Jalapeño

 OpenAI unveils its first custom chip, built by Broadcom | TechCrunch

OpenAI unveiled its first custom-built inference processor, designed and manufactured in collaboration with Broadcom. Named Jalapeño, the new processor was designed specifically for the unique needs of OpenAI’s inference systems. OpenAI’s own AI models assisted in the development of the chip, the company said. While the chip is still being tested, OpenAI says early results show significantly better performance-per-watt than current state-of-the-art alternatives.

OpenAI x Broadcom — The OpenAI Podcast Ep. 8 - YouTube

Net Zero vs. Passive House

 Net Zero vs. Passive House Clarified - EP #2 - YouTube @EkoBuilt Passive Homes

Designer builds efficient off-grid Passive House in Colorado - YouTube

Exploring Passive House Design - 90% Energy Savings! - YouTube
Undecided with Matt Ferrell

Numbers & Facts: Passive House vs Code-Built Homes - YouTube

This video from EkoBuilt Passive Homes compares code-built homes to passive houses, highlighting significant differences in insulation, efficiency, and cost.

Key Comparison Points:

• Building Envelope & Insulation (R-Values): Passive houses utilize superior insulation to maintain a stable environment. While code-built homes often rely on standard fiberglass or rock wool, passive houses use materials like cellulose (which is 100% recycled/renewable and can be made fire/mold-proof with borate) to reach significantly higher R-values (0:56 - 5:51).
  • Walls: R-24 (code) vs. R-75 (passive)
  • Floors: R-10 (code) vs. R-40 (passive)
  • Windows/Doors: R-3 (code) vs. R-12 (passive)
  • Roofs: R-36 (code) vs. R-110 (passive)
• Window Performance: Passive house windows are triple-glazed, thermal bridge protected, and engineered to maximize solar heat gain in winter (6:02 - 8:05).
• Cost & Savings:
• Incremental Build Cost: Building a passive house is roughly 5-10% more expensive initially due to the high-performance envelope (2:19).
• Utility Savings: While the average Canadian utility bill is ~50/month** (with net metering) or $100/month without it (8:20 - 10:28).

Energy-Efficient House Design | DIY Passive Home Build Walkthrough - YouTube byEkoBuilt Passive Homes

What Is A Passive House And Why You Need It (4K) - YouTube

Passive homes have gained significant recognition in recent years for their innovative and sustainable design principles. These houses are not just another architectural trend, but rather a response to the pressing need for energy-efficient and eco-friendly living spaces. With an emphasis on reducing energy consumption, passive houses have become a beacon of hope for mitigating the environmental impact of residential buildings.

Passive House - Fresh Air System Install & Costs (Zehnder) - YouTube

The ultimate guide to passive home design - YouTube

Passive House with sunroom

Passive House Walkthrough with Real Estate Agent Jen Stewart - YouTube

AI Model: GLM 5.2

Socialists Sweep NYC, China Catches Up in Coding, AI Memory Crunch, Micron's Blowout Quarter - YouTube @All-In Podcast

The Chinese AI model highlighted in the video as being highly capable and competitive with U.S. frontier models is GLM 5.2, released by Z.AI (45:18).

Key details about this model include:

• Open Source: It is released under the MIT license, meaning it is free to download, self-host, and modify (45:26 - 45:58).
• Technical Scale: It features 744 billion parameters and a 1 million token context window (45:26).
• Performance: It has achieved significant benchmarks, notably scoring 51 points on the Artificial Analysis intelligence index—the highest for any open-weight model—and performing very competitively against models like GPT 5.5 and Claude Opus 4.8 (46:06 - 46:22).
• Efficiency: The model is noted for being significantly cheaper to run (85% cheaper than GPT 5.5) and optimized for Huawei hardware (46:23, 59:33).

GLM 5.1 Thinks Strategically, Data-Center Revolt Intensifies, When Helpful LLMs Turn Unhelpful, Humanoid Robots Get to Work @DeepLearningAi (Andrew Ng)

GLM 5.2 API & Playground | Fireworks AI

GLM 5.2 - API Pricing & Benchmarks | OpenRouter

zai-org/GLM-5.2 · Hugging Face

visual: hexagonal world map

 Delivering a dynamic hexagonal world map in just 10kb | Calibre Blog

Awesome GoLang frameworks, libraries, and software

a very comprehensive list of useful Go tools

Awesome Go A curated list of awesome Go frameworks, libraries and software

AI: /goal Clone Excel, full feature parity

predictable, predicted, irrelevant?

Would this AI create many Office clone apps? Unlikely. 

There is already free OpenOffice, why bother creating more of the "same".
Why not improve, come with something new and better? That would be useful and interesting!

But the Microsoft stock price is falling, as well as Salesforce, Oracle,
possibly for some other reasons...

 AI can run for days on its own now - YouTube by @matthew_berman

msft stock price - Google Search

crm stock price - Google Search

orcl stock price - Google Search

on the other side, "AI first" Google is going strong

goog stock price - Google Search

ideas: DIY Concrete Block & Wood Steps

only gravity!

 DIY Concrete Block & Wood Steps: A Purposeful Backyard Ascent! ✨🪵 - YouTube

React developers => TanStack

40 min interview with creator of TanStack

Why React Developers Are Leaving Next.js for TanStack - YouTube


TanStack | The open-source application stack for the web.
The open-source application stack for the web.

Headless, type-safe, composable tools for building modern web applications that work naturally for developers and reliably for agents.

Is TanStack Start better than Next.js? - YouTube by MaxS

A totally subjective comparison of the most used and most hyped (?) React fullstack frameworks.

I moved off of Next.js - YouTube by Theo - t3.gg

I've officially moved T3 Chat off of NextJS and onto TanStack Start.

Neon DB + AI: psql C => TypeScript

Why Neon? - Neon Docs

Neon Postgres is a fully managed, serverless PostgreSQL database designed for modern applications. It separates the compute engine from the storage layer, allowing the database to scale up instantly to handle traffic spikes, scale down to zero when idle to save costs, and offer unique "branching" capabilities. [1, 2, 3, 4, 5]

Key Features

• Database Branching: Functions like Git for your data. You can instantly create an isolated, copy-on-write clone of your production database (schema and data) for testing, staging, or CI/CD pipelines.
• Serverless Auto-Scaling: Automatically adjusts compute resources (CPU and memory) to match your workload demands without requiring manual provisioning.
• Scale-to-Zero: When your database is completely inactive, it suspends itself, ensuring you only pay for storage rather than idle compute time.
• Point-in-Time Recovery: You can instantly restore your database to the exact state it was in at any specific second, preventing data loss during bad migrations. [4]

How It Works

Instead of tying your database to a single virtual machine with a fixed amount of storage, Neon decomposes the architecture:

1. Compute Layer: Stateless PostgreSQL nodes that handle your queries. Because they hold no permanent data, they can spin up and down on demand.
2. Storage Layer: A custom, multi-tenant storage engine that persists the transaction logs and data. This separation is what enables instant branching and zero-downtime scaling. [6, 8, 9]

Why Developers Use It

• Cost-Effective for Startups: It offers a highly capable free tier, and because you only pay for what you use, it is a budget-friendly option for side projects and small teams.
• Works with Serverless Architectures: Neon provides serverless drivers allowing you to query the database over HTTP or WebSockets, which is ideal for modern serverless functions or edge environments (like Cloudflare Workers) that cannot maintain traditional TCP connections.
• Full Postgres Compatibility: It supports popular Postgres extensions like (for AI and vector embeddings) and PostGIS. [13]

You can manage your projects or spin up a free database instantly through the Neon Console. For a deeper dive into how their engine works, you can view the Neon GitHub Repository. [2, 4, 5]

[1] https://github.com/neondatabase/neon
[2] https://cevheri.medium.com/neon-the-next-generation-opensource-serverless-postgresql-ab86e367627f

[3] https://www.youtube.com/watch?v=7KglfRnIlfc

[4] https://www.youtube.com/watch?v=wLAeC-r4kfE

[5] https://www.readysetcloud.io/blog/allen.helton/serverless-postgres-with-neon/

[6] https://neon.com/

[7] https://encore.dev/docs/platform/infrastructure/neon

[8] https://semaphore.io/blog/neon-database

[9] https://neon.com/docs/introduction/architecture-overview

[10] https://www.reddit.com/r/Database/comments/1pfqlix/is_neontech_postgresql_good_for_small_startup/

[11] https://www.dbpro.app/blog/neon-vs-supabase

[12] https://www.youtube.com/watch?v=EB0Nu_e9wCs

[13] https://neon.com/pricing

What is Neon? - YouTube

Introducing neon.ts: infrastructure as code for your Neon projects - Neon

neondatabase/neon: Neon: Serverless Postgres. We separated storage and compute to offer autoscaling, code-like database branching, and scale to zero. @GitHub

Shipping psql without psql: a pure-TypeScript client in neonctl - Neon

Here is a quick summary of the article:

Shipping psql without psql details how Neon embedded a pure-TypeScript psql client directly into their neonctl CLI so users don't have to worry about missing native Postgres dependencies on their systems.

Key Takeaways:

• AI-Generated Code: The ~24,000 lines of code for the client were re-written from C to TypeScript almost entirely by an AI coding agent (Claude Code).

• Rigorous Validation: Because the tool handles sensitive auth and TLS paths, the engineering team couldn't just trust the AI's output. They built a strict conformance harness to verify byte-exact parity with PostgreSQL's official regression tests.

• Adversarial Review: Human engineers conducted deep security reviews, catching subtle vulnerabilities in the AI's "correct-looking" code.

• The Lesson: AI is highly capable of reimplementing complex tools, but true reliability comes entirely from the human-built testing harnesses and security reviews that validate the output.

neonctl psql.ts vs psqlsh.ts?

It is completely fair to find this confusing! At first glance, it looks like Neon built the exact same thing twice.

While both are TypeScript-based emulations of psql, they were built at different times to solve two completely different problems, run in completely different environments, and have vastly different levels of accuracy.

Here is the breakdown of why they both exist and how they differ:

Quick Comparison

Feature	psql.sh (The Web Playground)	neonctl psql (The Embedded CLI)
Where it runs	In your web browser (Chrome, Safari, Firefox)	On your local machine (Terminal via Node.js)
Primary Purpose	A frictionless marketing tool & playground. Lets anyone try Postgres in 2 seconds without opening a terminal.	A local safety net. Prevents command not found: psql errors when a developer doesn't have Postgres installed.
Connection Type	WebSockets (Browsers cannot open raw TCP database connections)	Standard TCP Sockets + TLS
Accuracy Scope	Lightweight approximation. Implements the most common \d commands.	Byte-perfect emulation. Matches ~24,000 lines of C code and passes official PostgreSQL regression tests.
Bonus Features	Built-in AI generation (\ai create a users table).	Full support for complex local files like .psqlrc and ~/.pgpass.

---

The Deep Dive: Why have both?

1. psql.sh is a Web Toy & Interactive Demo

Before Neon worried about local CLI dependencies, they wanted a way to showcase how fast their serverless database branches could boot.

• The Goal: If you are reading a blog post or documentation, they want you to click a link and immediately start typing SQL.

• The Limit: Web browsers strictly forbid raw TCP networking (the protocol Postgres speaks). Because of this, psql.sh is highly specialized: it talks over WebSockets to a proxy, uses a lighter command translation library, and throws in fun web features like an AI SQL assistant. You would never use this to manage a heavy production database.

2. neonctl psql is a Heavy-Duty Local Fallback

This is the feature detailed in the Shipping psql without psql article.

• The Goal: You are a developer working locally in your terminal. You type neonctl psql to connect to your database, but you are on a brand new Mac or a stripped-down Docker container that doesn't have PostgreSQL installed. Instead of crashing and forcing you to go install Postgres, neonctl secretly activates this pure-TypeScript engine.

• The Engineering: Because this is for local, professional developers, it cannot just be a "lightweight approximation" like the web version. It has to support every edge case—complex security certificates, massive database schema formatting, and secure authentication loops. This is the massive, Claude-generated engine that required strict byte-for-byte conformance testing against Postgres's actual C source code.

Summary

Think of psql.sh as a lightweight, web-optimized "preview" for quick experimentation, and the neonctl embedded psql as a bulletproof, local "drop-in replacement" designed to keep your terminal workflows from breaking.