← Ch 2 · Contents · ↑ Contents
From self-1.md (Oct 2024 – Feb 2026)
AI Coding Tools & Agents
Thoughts On A Month With Devin answer.ai — https://www.answer.ai/posts/2025-01-08-devin.html
Answer.ai’s honest retrospective after a month using Devin (Cognition’s AI software engineer agent) in real work. Finds Devin impressive for well-scoped isolated tasks but unreliable for complex multi-file refactoring or tasks requiring judgment about existing codebase conventions. [→ machine-learning-ai; infrastructure-devops]
GitHub Copilot for the Command Line christianheilmann.com — https://christianheilmann.com/2023/03/29/github-copilot-for-the-command-line-is-amazing/
Christian Heilmann’s enthusiasm for GitHub Copilot in the command line — demonstrating how natural language shell assistance reduces the cognitive load of remembering complex command syntax and flags. [→ infrastructure-devops; machine-learning-ai]
Create Apps Without Writing Any Code: DeepSeek-R1 and RooCode geeky-gadgets.com — https://www.geeky-gadgets.com/no-code-app-creation-guide/
A tutorial on using RooCode (an agentic coding assistant for VSCode) with DeepSeek-R1 as the backend to generate complete applications from natural language descriptions, demonstrating the no-code frontier of AI-assisted software development. [→ machine-learning-ai; infrastructure-devops]
GitHub - RooVetGit/Roo-Code github.com — https://github.com/RooVetGit/Roo-Code
Roo-Code is an open-source agentic coding assistant for VSCode — a fork of Cline with additional features including custom model backends, persistent memory, and workflow automation. Supports DeepSeek, Claude, and other LLMs as the reasoning engine. [→ machine-learning-ai; infrastructure-devops]
Julian Goldie: Deep Seek R1 and RooCode x.com — https://x.com/JulianGoldieSEO/status/1884960230804783591
A tweet demonstrating how to use DeepSeek-R1 (a low-cost reasoning model) with RooCode to build complete web applications at minimal cost — part of the wave of “local AI coding” content following DeepSeek’s release. [→ machine-learning-ai]
Cline: DeepSeek R1 as ‘code archaeologist’ x.com — https://x.com/thankscline/status/1883201442682720700
Cline’s tweet showing DeepSeek-R1 being used as a “code archaeologist” — using the model’s reasoning capabilities to navigate and understand large existing codebases rather than generating new code from scratch. [→ machine-learning-ai; infrastructure-devops]
John Rush: Browser-Use tools review x.com — https://x.com/johnrushx/status/1883872256121774401
John Rush’s evaluation of browser automation tools that allow LLMs to control a web browser — including Browser-Use, Playwright-based agents, and OpenAI’s Operator. Compares reliability, cost, and task completion rates. [→ machine-learning-ai; infrastructure-devops]
I replaced CoPilot AI with DeepSeek androidauthority.com — https://www.androidauthority.com/deepseek-vs-copilot-3520404/
An Android Authority comparison of DeepSeek vs Microsoft Copilot for daily coding assistance — finding DeepSeek competitive on quality at a fraction of the cost, contributing to discussion of the “DeepSeek moment” in the AI industry. [→ machine-learning-ai]
Tech with Mak: The person who built Claude Code just mass-leaked the thinking behind it 45 minutes of design decisions, mistakes, and where it’s all going — x.com — https://x.com/techNmak/status/2024041443837526375?s=20
A tweet pointing to a 45-minute talk by the Claude Code creator covering the design philosophy, technical decisions, early mistakes, and future direction of Claude Code as an agentic coding tool — one of the most candid insider accounts of AI coding agent design. [→ machine-learning-ai; infrastructure-devops]
Sentient: Anthropic free short course on Claude Code agents x.com — https://x.com/sentientt_media/status/2025142906051498085?s=20
A tweet announcing Anthropic’s free short course on building and deploying Claude Code agents and skills — covering the Claude Agent SDK, hooks, MCP servers, and workflow automation deployable across Claude Code, API, and VS Code. [→ machine-learning-ai; infrastructure-devops]
Paweł Huryn: Claude Code — The Ultimate Guide for PMs x.com — https://x.com/PawelHuryn/status/2025470280945041547?s=20
Paweł Huryn’s guide to Claude Code aimed at product managers — explaining how to use Claude Code for product discovery, spec writing, and light coding tasks, bridging the gap between PM workflows and AI-assisted development. [→ machine-learning-ai; infrastructure-devops]
Matthew Berman: OpenClaw (Claude Code) 21 use cases x.com — https://x.com/MatthewBerman/status/2023843493765157235?s=20
Matthew Berman’s thread listing 21 practical use cases for OpenClaw (a framework built on Claude Code) — ranging from automated PR review and code generation to data analysis pipelines and agent orchestration. [→ machine-learning-ai; infrastructure-devops]
GREG ISENBERG: How to use OpenClaw to spin up 24/7 digital employees x.com — https://x.com/gregisenberg/status/2024247983999521123?s=20
Greg Isenberg’s thread on building “24/7 digital employees” using OpenClaw — AI agents that continuously monitor, respond, and execute tasks autonomously, enabling small teams to punch above their weight with AI-assisted operations. [→ machine-learning-ai; infrastructure-devops]
Ziwen: How to Run a 24/7 AI Company with OpenClaw for $50/Month x.com — https://x.com/ziwenxu_/status/2023610499024171077?s=20
A practical cost breakdown of running continuous AI agent workflows on a Mac Mini using OpenClaw — showing that Claude API costs for autonomous agent operations can be kept under $50/month for a solo founder or small team. [→ machine-learning-ai; infrastructure-devops]
Rishabh: Full OpenClaw Course in 6 Hours x.com — https://x.com/Rixhabh__/status/2025844767624700193?s=20
A tweet linking to a 6-hour comprehensive OpenClaw course covering setup, agent design, workflow automation, and real-world use cases — providing structured learning for those wanting to build production Claude Code agentic systems. [→ machine-learning-ai; infrastructure-devops]
GitHub - duolingo/slack-mcp: OAuth-based multi-user Slack MCP server github.com — https://github.com/duolingo/slack-mcp
Duolingo’s open-source Slack MCP (Model Context Protocol) server with OAuth-based authentication and HTTP transport — enabling AI assistants like Claude to read and write Slack messages as a properly authenticated multi-user service rather than a single bot token. [→ infrastructure-devops; machine-learning-ai]
The Lobster Takeover: Developers Buying Mac Minis to Run Their Own AI Agents starryhope.com — https://www.starryhope.com/minipcs/clawdbot-mac-mini-ai-agent-trend/
An article on the trend of developers buying Mac Minis (M4 Pro) as always-on local AI agent servers — running Claude Code, OpenClaw, or other agents continuously at low power cost instead of paying cloud compute rates. [→ machine-learning-ai; infrastructure-devops]
Signal creator Moxie Marlinspike wants to do for AI what he did for messaging Introducing Confer, end-to-end AI assistant — arstechnica.com — https://arstechnica.com/security/2026/01/signal-creator-moxie-marlinspike-wants-to-do-for-ai-what-he-did-for-messaging/
Ars Technica on Moxie Marlinspike’s new AI assistant Confer — designed with end-to-end encryption and privacy-first architecture so the AI provider cannot read conversation content. Applies Signal’s messaging privacy model to LLM assistants. [→ machine-learning-ai; infrastructure-devops]
“SaaS is dying as a business category” As AI agents commoditize software — calcalistech.com — https://www.calcalistech.com/ctechnews/article/hjlvyl7lze
An Israeli tech media analysis arguing that AI agents are commoditising SaaS: when AI can perform tasks that previously required specialised software (CRM, project management, etc.), the moat of product-specific workflows erodes. [→ machine-learning-ai]
MCP (Model Context Protocol)
MCP introduction modelcontextprotocol.io — https://modelcontextprotocol.io/introduction
The official introduction to Anthropic’s Model Context Protocol (MCP) — an open standard for connecting LLMs to data sources and tools. MCP defines a client-server protocol where MCP servers expose capabilities (tools, resources, prompts) that LLM clients can discover and invoke. [→ machine-learning-ai; infrastructure-devops]
John Rush: What is MCP & why it’s a big deal x.com — https://x.com/johnrushx/status/1897655569101779201
John Rush’s layperson explanation of MCP’s significance: by standardising how LLMs connect to external tools, MCP enables an ecosystem where any AI application can use any MCP server — analogous to how USB standardised peripheral connections. [→ machine-learning-ai; infrastructure-devops]
Nix / Build Tools
Scala Native & Nix - a match made in heaven? https://youtu.be/old8N9u3QKU
A YouTube talk exploring the combination of Scala Native (Scala compiled to native code) and Nix (reproducible builds) — arguing that Nix’s hermetic builds plus Scala Native’s ahead-of-time compilation make for deterministic, fast, containerless Scala deployment. [→ scala-functional-programming; infrastructure-devops]
typelevel-nix github.com — https://github.com/typelevel/typelevel-nix
A Nix flake providing development environments for Typelevel Scala projects (Cats, CE3, http4s, etc.) — pre-configured with JDK, sbt, Scala CLI, and Metals LSP, enabling reproducible dev environments with a single nix develop command. [→ scala-functional-programming; infrastructure-devops]
Why you should flox every day (Nix/flox) etorreborre.blog — https://etorreborre.blog/why-you-should-flox-every-day
Eric Torreborre’s blog post on Flox — a user-friendly layer on top of Nix that simplifies managing developer environments while retaining Nix’s reproducibility guarantees. Argues daily use of Flox normalises environment-as-code and eliminates “works on my machine” issues. [→ infrastructure-devops; scala-functional-programming]
Rust, Zig & Systems Programming
[SEI’ 24] Modern Systems Programming: Rust and Zig https://youtu.be/4aLy6qjhHeo
A Software Engineering Institute 2024 talk comparing Rust and Zig for modern systems programming — covering safety guarantees, performance characteristics, comptime, and which language to choose for different embedded and systems contexts. [→ scala-functional-programming; infrastructure-devops]
Zig And Rust matklad.github.io — https://matklad.github.io/2023/03/26/zig-and-rust.html
Matklad’s (the main author of rust-analyzer) thoughtful comparison of Zig and Rust: Zig as “better C” with simplicity and comptime; Rust as safe systems programming with ownership — complementary tools for different problem classes. [→ scala-functional-programming; infrastructure-devops]
A new tool for visualizing Rust lifetimes https://youtu.be/NV6Xo_el_2o
A YouTube demo of a new IDE tool for visually rendering Rust lifetime annotations on code — helping developers understand borrow checker errors by showing lifetime scopes as overlapping regions rather than text annotations. [→ scala-functional-programming; infrastructure-devops]
C++ to Rust Phrasebook A book to help translate C++ idioms into Rust — cel.cs.brown.edu — https://cel.cs.brown.edu/crp/
A Brown University resource mapping C++ idioms to their Rust equivalents — covering RAII, ownership, templates vs generics, exception handling vs Result, and move semantics. Useful for C++ programmers transitioning to Rust. [→ scala-functional-programming; infrastructure-devops]
Forrest: Rust replacing Java and Scala (xAI x-algorithm repo) x.com — https://x.com/ForrestPKnight/status/2018768871474315761?s=20
A tweet noting that xAI’s x-algorithm repository (the core algorithm powering X/Twitter) is written in Rust — where Java or Scala might have been expected — suggesting Rust’s performance and safety advantages are drawing adoption even in traditionally JVM-based domains. [→ scala-functional-programming; infrastructure-devops]
After Three Years, Modular’s CUDA Alternative Is Ready eetimes.com — https://www.eetimes.com/after-three-years-modulars-cuda-alternative-is-ready/
EE Times reports on Modular’s Mojo and MAX platform reaching production readiness — offering a CUDA alternative that runs on diverse hardware (NVIDIA, AMD, Apple Silicon) using a portable compilation stack. [→ machine-learning-ai; infrastructure-devops]
Modular: DeepSeek’s Impact on AI (Democratizing AI Compute, Part 1) modular.com — https://www.modular.com/blog/democratizing-compute-part-1-deepseeks-impact-on-ai
Modular’s analysis of how DeepSeek’s efficient training methodology and open-weight release is democratising AI compute — enabling capable models to run on consumer hardware and reducing the cost barrier for AI deployment. [→ machine-learning-ai; infrastructure-devops]
Terminal & Dev Environment
Ghostty is Probably The Best Terminal Emulator https://youtu.be/3wq0RFAvNo
A YouTube review of Ghostty — Mitchell Hashimoto’s new GPU-accelerated terminal emulator for macOS and Linux. Combines the speed of Alacritty with the features of Kitty and better macOS integration, using a native rendering engine rather than web technologies. [→ infrastructure-devops]
Programming in the Sun: A Year with the Daylight Computer wickstrom.tech — https://wickstrom.tech/2025-10-10-programming-in-the-sun-a-year-with-the-daylight-computer.html
Oskar Wickström’s year-long review of the Daylight Computer — an e-ink/LCD hybrid display device designed for outdoor use and reduced eye strain. Evaluates it as a daily programming machine: editing, terminals, and the experience of a sunlight-readable developer environment. [→ infrastructure-devops]
AWS / Cloud / Kubernetes
10 Tips for AWS Lambda functions performant medium.com — https://medium.com/@jake.bazin/10-tips-for-keeping-your-aws-lambda-functions-performant-f0504949659a
A Medium article covering Lambda performance optimisation: cold start reduction (provisioned concurrency, container reuse), memory tuning, connection pooling, efficient serialisation, and avoiding synchronous invocations for long-running tasks. [→ infrastructure-devops]
AI experts return from China stunned: The U.S. grid is so weak, the race may already be over Data center infrastructure comparison — fortune.com — https://fortune.com/2025/08/14/data-centers-china-grid-us-infrastructure/
Fortune’s alarming report from AI researchers who visited Chinese data center facilities: China’s electrical grid infrastructure can support much larger AI training clusters than equivalent US facilities — raising concerns about long-term AI compute capacity parity. [→ machine-learning-ai; infrastructure-devops]
Kubernetes for Scala developers rockthejvm.com — https://rockthejvm.com/articles/kubernetes-for-scala-developers
Rock the JVM’s guide to Kubernetes from a Scala developer’s perspective — covering Pod, Deployment, Service, and ConfigMap concepts with Scala application examples, Helm chart basics, and deployment patterns relevant to Akka/ZIO services. [→ scala-functional-programming; infrastructure-devops]
CNCF Tech Radar Report (Nov 2025) cncf.io — https://www.cncf.io/wp-content/uploads/2025/11/cncf_report_techradar_111025a.pdf
The Cloud Native Computing Foundation’s November 2025 Technology Radar — surveying cloud native practitioners on Adopt/Trial/Assess/Hold recommendations for tools across observability, platforms, databases, and AI infrastructure. [→ infrastructure-devops]
Electric Vehicles & E-Bikes
GT73 Electric Dirt Bike 2000W amazon.com — https://www.amazon.com/GT73-Electric-Motorcycle-Hydraulic-Suspension/dp/B0DLKY63Q8
An Amazon product listing for the GT73 electric dirt bike — a 2000W off-road electric motorcycle with hydraulic suspension. Noted as a relatively affordable electric off-road option. [→ health-fitness]
Rivian’s ALSO E-Bike The Coolest E-Bike I’ve Ever Seen https://youtu.be/K_vBM5eY_vk
A YouTube review of the ALSO TM-B e-bike from Rivian’s micromobility startup — featuring full suspension, a mid-drive motor, and trail-focused geometry. Hailed as unlike any other e-bike due to its mountain bike DNA and Rivian’s quality pedigree. [→ health-fitness]
ALSO TM-B In-Depth ($4,500 NextGen Rivian eBike) https://youtu.be/9004qAghecM
DC Rainmaker’s in-depth review of the ALSO TM-B — Rivian’s premium full-suspension e-bike at $4,500. Covers motor performance, battery life, suspension behaviour, and whether the price premium over conventional e-bikes is justified. [→ health-fitness]
Rivian’s full-suspension e-bike is like nothing we’ve seen before singletracks.com — https://www.singletracks.com/mtb-gear/the-also-tm-b-is-a-full-suspension-e-bike-unlike-any-other-youve-seen/
Singletracks (mountain bike media) review of the ALSO TM-B from a trail riding perspective — evaluating how the full-suspension design handles technical terrain versus conventional hardtail e-MTBs. [→ health-fitness]
Review: I rode the ALSO TM-B E-Bike: Rivian Micromobility Startup’s First EV rivianforums.com — https://www.rivianforums.com/forum/threads/review-i-rode-the-also-tm-b-e-bike-rivian-micromobility-startup%E2%80%99s-first-ev.51873/
A Rivian owner community review of the ALSO TM-B e-bike — providing a user perspective from the Rivian enthusiast community, including ride impressions and comparison to other premium e-bikes. [→ health-fitness]
Gearless Magnet Bike https://youtu.be/Dg8oVR4k5Dk?si=7lDvJArmgqjV4Ilo
A YouTube video on a gearless bicycle using a magnetic drive system — eliminating the chain and derailleur entirely. Explores how magnetic coupling can transmit power from pedals to wheel without mechanical contact. [→ health-fitness]
The Insane Engineering Of A Land Speed Record Bike https://youtu.be/e97w4P0LB0Q?si=BueUijDM0D26mKu-
Real Engineering’s breakdown of the engineering behind land speed record bicycles — aerodynamic fairings, recumbent positions, gearing ratios, and how cyclists achieve speeds over 280 km/h (175 mph) with motor pacing. [→ health-fitness]
Bambuk E-Trike Tandem Recumbent Trike rad-innovations.com — https://www.rad-innovations.com/bambuk-tandem.html
The Bambuk is a tandem electric recumbent trike — a low-slung, fully faired cargo-capable vehicle for two riders. An unusual category combining the efficiency of recumbent cycling with electric assist and tandem capacity. [→ health-fitness]
Schwinn IC4 Review (2025) yourexercisebike.com — https://yourexercisebike.com/schwinn-ic4-review/
A comprehensive review of the Schwinn IC4 indoor cycling bike — covering resistance feel, Bluetooth connectivity (pairs with Peloton, Zwift apps), build quality, and value comparison against the Peloton Bike at roughly one-third the price. [→ health-fitness]
Upgrade Your Indoor Bike with SmartSpin2K V3 yourexercisebike.com — https://yourexercisebike.com/smartspin2k-guide/
The SmartSpin2K is an aftermarket smart resistance controller that can be retrofitted onto existing spin bikes — adding Bluetooth ERG mode, Zwift compatibility, and automatic resistance control without buying a new smart trainer. [→ health-fitness]
GitHub - Quinus/WhooshConnect: App that uploads MyWhoosh activities to Garmin Connect github.com — https://github.com/Quinus/WhooshConnect
An open-source tool for automatically uploading indoor cycling activities from the MyWhoosh virtual cycling platform to Garmin Connect — solving the friction of multi-platform fitness data consolidation. [→ health-fitness; infrastructure-devops]
Home Gym Equipment
Saris H3 Smart Trainer In-Depth Review dcrainmaker.com — https://www.dcrainmaker.com/2019/08/saris-trainer-review.html
DC Rainmaker’s thorough review of the Saris H3 direct-drive smart trainer — a Zwift-compatible wheel-off trainer with accurate power measurement, quiet flywheel, and strong ERG mode. [→ health-fitness]
Rogue Echo Bike Review morningchalkup.com — https://morningchalkup.com/2022/06/23/rogue-echo-bike-review/
Morning Chalk Up’s review of the Rogue Echo Bike — the CrossFit community’s favourite air resistance assault bike. Covers build quality, resistance feel, console, and why it’s considered the best-in-class for high-intensity conditioning work. [→ health-fitness]
Schwinn’s Airdyne AD7 vs Rogue’s Echo Bike bikevsbike.com — https://bikevsbike.com/schwinns-airdyne-ad7-vs-rogues-echo-bike/
A head-to-head comparison of two iconic air bikes: the Schwinn Airdyne AD7 (the original classic) vs the Rogue Echo Bike (the modern CrossFit standard). Covers resistance curve, durability, seat comfort, and price. [→ health-fitness]
REP Strive Series Curved Treadmill repfitness.com — https://repfitness.com/products/rep-strive-series-curved-treadmill?variant=43288603099294
REP Fitness’s non-motorised curved treadmill — a manual treadmill where the belt speed responds to the runner’s position on the curved surface. Promotes natural running gait and higher metabolic cost than motorised treadmills. [→ health-fitness]
Expert-Tested: Torque Relentless Ripper Review (2025) — SkiErg alternative garagegymreviews.com — https://www.garagegymreviews.com/torque-relentless-ripper-review
Garage Gym Reviews’ evaluation of the Torque Relentless Ripper — an air-resistance pull-down machine designed as a SkiErg alternative. Tests whether it replicates the cardiovascular and upper body benefits of the Concept2 SkiErg. [→ health-fitness]
The Most Underrated Conditioning Equipment for Home Gyms (Concept 2 SkiErg) https://youtu.be/BuhTVvLfCk4
A YouTube video making the case for the Concept2 SkiErg as the most underrated home gym conditioning tool — comparing it to rowing machines and assault bikes for cardiovascular training, lat engagement, and core activation. [→ health-fitness]
Rowing Machine Vs. SkiErg: Which Is Better For Conditioning? fitnessprogramer.com — https://fitnessprogramer.com/rowing-machine-vs-skierg/
A structured comparison of indoor rowing and SkiErg for cardiovascular conditioning — muscle groups activated, caloric burn, impact on lower vs upper body, and which to choose for different fitness goals. [→ health-fitness]
Air Bike Vs Stationary Bike: What’s the Difference? garagegymreviews.com — https://www.garagegymreviews.com/air-bike-vs-stationary-bike
Garage Gym Reviews’ comparison of fan-powered air bikes (Rogue Echo, Assault) versus magnetic resistance stationary bikes — covering resistance feel, upper body engagement, caloric burn rate, and suitability for HIIT vs steady-state cardio. [→ health-fitness]
REP AB-3000 2.0 FID Adjustable Weight Bench repfitness.com — https://repfitness.com/products/ab-3000-fid-adjustable-bench?variant=42149286772894
REP Fitness’s flat-incline-decline adjustable bench — a home gym staple for chest, shoulder, and back pressing exercises. The AB-3000 2.0 is well-regarded for stability and adjustability at a lower price than commercial benches. [→ health-fitness]
WORLDS First ALL ELECTRIC Motorhome (Coachmen RVEX 2026) https://youtu.be/0x_ZCLjNnWE
A YouTube video on the Coachmen RVEX 2026 — claimed as the world’s first all-electric motorhome, running on a large battery pack without a combustion engine or generator. Covers range, charging infrastructure, and the practical challenges of electric RV travel. [→ miscellaneous]
Garmin & Health Wearables
How does Garmin measure stress, and is it really accurate? digitaltrends.com — https://www.digitaltrends.com/mobile/how-does-garmin-measure-stress/
Digital Trends’ investigation into Garmin’s stress monitoring — explaining the HRV-based algorithm (Heart Rate Variability derived from GPS watch sensors), its accuracy compared to clinical HRV measurements, and the limitations of wrist-based stress detection. [→ health-fitness]
Why Washington’s all-in on smart rings (Oura) politico.com — https://www.politico.com/news/2026/02/09/oura-ring-lobbying-rfk-maha-washington-00770320
Politico’s report on Oura Ring’s successful lobbying efforts in Washington — connecting the health wearable company’s growth to the MAHA (Make America Healthy Again) movement and how wearable health data is entering policy discussions. [→ health-fitness; miscellaneous]
Body Pod by Hume Health: In-Home Body Composition Analysis humehealth.com — https://humehealth.com/pages/hume-body-pod
The Hume Body Pod is a consumer air displacement plethysmography (ADP) device — the same technology used in clinical BOD POD machines — enabling accurate body composition measurement (fat vs lean mass) at home without a clinic visit. [→ health-fitness]