Blog
Essays on building judgment-heavy systems.
When an engineer intentionally breaks what works, most teams brace for the firefight.
When an engineer intentionally breaks what works, most teams brace for the firefight. Last week Josh Fathi did exactly that, publishing the results under a Hac…
Metabolic Syndrome Beyond GLP-1s: What Peptide Research Says About Insulin Sensitivity
Metabolic Syndrome Beyond GLP-1s: What Peptide Research Says About Insulin Sensitivity This article is for education only and does not constitute medical advic…
⏱ Timeout — denying command
⏱ Timeout — denying command ⏱ Timeout — denying command On May 24, our pipeline reported Reddit comments and arXiv as "fresh" by volume count. They were not fr…
Cognitive Peptides Explained: Semax, Selank, and the Research on Focus & Stress
Cognitive Peptides Explained: Semax, Selank, and the Research on Focus & Stress Most nootropic content on the internet falls into two camps: breathless biohack…
The Sleep & Recovery Stack: How DSIP, BPC-157, and Glutathione Work Together
The Sleep & Recovery Stack: How DSIP, BPC-157, and Glutathione Work Together This article is for education only and does not constitute medical advice. It is n…
AI Video Pipeline: How We Cut Costs From $0.52 to $0.08 Per Clip
AI Video Pipeline: How We Cut Costs From $0.52 to $0.08 Per Clip TL;DR: Tacavar's video pipeline generates production-ready short-form clips at $0.08 each by r…
The Reality: Patients Are Already Using AI for Health Questions
The Reality: Patients Are Already Using AI for Health Questions The shift happened fast. In 2023, most patients had never used a chatbot for medical advice. By…
How LuxeFit Classifies Evidence
How LuxeFit Classifies Evidence Before we discuss specific compounds, here is the framework we use internally when reviewing research for patient education: |…
Why Every AI Holding Company Needs an Agent Operating System
Why Every AI Holding Company Needs an Agent Operating System Most AI holding companies will fail for the same reason most venture studios fail: they confuse ha…
Multi-Agent Trading Frameworks Are Surging — Here's How They Compare
Multi-Agent Trading Frameworks Are Surging — Here's How They Compare A research group at UCLA and MIT shipped TradingAgents in late 2024. Six months later it h…
The Founder's AI Stack 2026: 12 Tools We Actually Use at Tacavar
The Founder's AI Stack 2026: 12 Tools We Actually Use at Tacavar The fastest way to waste money on AI is to buy twelve tools before you know which one owns whi…
Judgment Compounds: The Tacavar Framework for AI-First Decisions
Judgment Compounds: The Tacavar Framework for AI-First Decisions The cron jobs were “healthy.” The work was still failing. That was one of the more useful Taca…
What Is an AI Holding Company? (And Why the Model Beats VC for Operators)
What Is an AI Holding Company? (And Why the Model Beats VC for Operators) Most founders who raise venture capital learn the same lesson too late: the money is…
Why Agent Routing Matters More Than Prompt Engineering in Production AI
Why Agent Routing Matters More Than Prompt Engineering in Production AI The model said it understood, then it spent forty minutes inventing its way through the…
The Single Best Macro Signal for Crypto Trading (It's Free and Takes 10 Lines)
The cleanest leading signal for Bitcoin comes from the Federal Reserve's open API. It is not a proprietary on-chain metric. It is not a paid sentiment feed. It…
My Grafana Dashboard Looked Beautiful and Had Exactly One Operation in It
59 of 59 traces were the same 250ms heartbeat. The cron polling its own queue. I had just finished standing up Tacavar's observability stack: Tempo for distrib…
How One Server Can Run Whitelisted Commands on Another Without Sharing SSH Keys
Sharing root SSH keys was the obvious answer. I picked something smarter. At Tacavar, we run a two-droplet setup. The primary handles production traffic, Caddy…
Why LLM Routing Fails When Mid-Tier Models Pretend They Understand the Task
The real agent bug wasn't prompting. It was letting the wrong model decide. The hidden risk in autonomous agent routing Most AI agent architectures treat LLM r…
At 12:04 UTC on Friday, the self-heal cron logged its 24th run of the week.
At 12:04 UTC on Friday, the self-heal cron logged its 24th run of the week. Zero stuck runs. All Docker containers up. The Goose worker on qwen3-coder-plus ans…
Tirzepatide Cash-Pay Online: Safety, Screening, and Follow-Up
Tirzepatide Cash-Pay Online: Safety, Screening, and Follow-Up Tirzepatide has become one of the most discussed medications in cash-pay wellness and weight mana…
⚠ Compression model (gpt-5.2-codex) context is 400,000 tokens, but the main model's compression thre
⚠ Compression model (gpt-5.2-codex) context is 400,000 tokens, but the main model's compression threshold was 525,000 tokens. Auto-lowered this session's thres…
Glutathione and Wellness: What to Know Before a Cash-Pay Consult
Glutathione and Wellness: What to Know Before a Cash-Pay Consult Meta Description Glutathione is a key antioxidant made by the body. Learn what patients should…
How GHK-Cu signals the body to rebuild collagen — and what the research actually shows
How GHK-Cu signals the body to rebuild collagen — and what the research actually shows GHK-Cu (copper peptide) is among the best-studied peptides in the skin-r…
Retatrutide vs. Tirzepatide vs. Semaglutide: Cash-Pay GLP-1 Options in 2026
Retatrutide vs. Tirzepatide vs. Semaglutide: Cash-Pay GLP-1 Options in 2026 For patients exploring medical weight management in 2026, the question is no longer…
The week’s most useful failure didn’t happen in a dashboard.
The week’s most useful failure didn’t happen in a dashboard. It happened the morning after a clean bill of health. Twenty-four cron jobs had already been audit…
'Beautiful Dashboards, Empty Truth: The Observability Trap That Ships Blind Systems'
A perfect dashboard can hide a completely non-working system. Founder/operators at Tacavar.com do not have the luxury of mistaking presentation for proof. Josh…
Two Free Macro Signals We’d Track Before Paying for Another Trading Dashboard
Some of the best crypto signals on earth are still free. That is not a contrarian slogan. It is a practical operating principle Tacavar uses when deciding what…
How Tacavar Built Cross-Server Command Dispatch Without Sharing Root SSH Keys
Shared SSH keys are easy. Least-privilege automation is better. That tradeoff sits at the center of how Tacavar approaches autonomous infrastructure. When one…
Week 3: Cluster Detection Holds, Regime Shift
Week 3: Cluster Detection Holds, Regime Shift The cluster fix from week 2 held under real pressure. The regime flipped mid-week. Equity curve tracker live. +$2…
Automated Crypto Portfolio Management: From Manual to Autonomous
Automated Crypto Portfolio Management: From Manual to Autonomous Manual crypto portfolio management is a second job. Here's how automated rebalancing, risk con…
AI Crypto Trading Bot 2026: What Actually Works (And What Doesn't)
AI Crypto Trading Bot 2026: What Actually Works (And What Doesn't) Most AI trading bots overpromise and underdeliver. Here's an honest breakdown of what separa…
Why Our Critic Agent Vetoes Bad Trades
Why Our Critic Agent Vetoes Bad Trades Inside our adversarial risk architecture: how the critic agent blocks bad trades before execution, the veto conditions t…
How to Build an AI Trading Bot That Actually Works
How to Build an AI Trading Bot That Actually Works Most trading bots fail before they place a live trade — not because the strategy was wrong, but because the…
Prediction Market Trading Bot: How We Trade Polymarket with AI
Prediction Market Trading Bot: How We Trade Polymarket with AI Prediction markets are one of the sharpest alpha sources in 2026. Here's how we built an AI bot…
Best AI Crypto Trading Bots in 2026: Honest Comparison
Best AI Crypto Trading Bots in 2026: Honest Comparison We reviewed 8 AI crypto trading bots — 3Commas, Cryptohopper, Pionex, HaasOnline, Coinrule, Bitsgap, Shr…
Inside Our $10K Paper Trading Bot: 30 Days of Real Data
Inside Our $10K Paper Trading Bot: 30 Days of Real Data Twenty-five trades. Multiple strategies. LLM-driven decisions. Here's exactly what happened — the wins,…
OralMind: AI-Powered Dental Workflow SaaS
OralMind: AI-Powered Dental Workflow SaaS Introducing OralMind — an AI dental workflow platform that helps practitioners catch problems earlier, document faste…
Building Healthcare AI That Doctors Actually Use
Building Healthcare AI That Doctors Actually Use Most healthcare AI never makes it out of the lab. Here's how we're deploying AI into real dental and medical p…
How We Do SEO in 2026: A 4-Vertical Case Study
How We Do SEO in 2026: A 4-Vertical Case Study Google's SGE, AI content floods, and algorithm updates changed everything. Here's our 4-pillar framework for ran…
The Hidden Supply Chain Behind America's Biologics Boom
The Hidden Supply Chain Behind America's Biologics Boom The US biologics market will hit $500B by 2030. Here's the hidden supply chain powering America's biolo…
Week 2: First Trades, First Lessons
Week 2: First Trades, First Lessons The bot executed its first 8 paper trades. 62% win rate. An overtrading incident caught on Day 11 and fixed by Thursday. Fu…
Week 1: Setting the Baseline
Week 1: Setting the Baseline Initial setup complete. 9 strategies calibrated. First signals firing. Here's what happened in our first 7 days — including why we…
How We Build Our Trading Bot in the Open
How We Build Our Trading Bot in the Open A transparent look at our algorithmic trading system — paper trading crypto and Polymarket with 9 strategies, LLM-augm…
claude-max-proxy-loophole-closed-agent-sdk-migration
-|------|------| | Orchestration | Claude Agent SDK + custom skills | API billing | | Web browsing | browserbase/skills | Browserbase usage | | Heavy automatio…