Technology On the Net

Multi-step LLM agents do not just answer questions. They book flights, create tickets, charge cards, send emails, and write to databases. When step three fails after steps one and two already succeeded, you cannot roll everything back with a single database transaction. The tools live in different systems, each with its own API and its own idea of "done." This post covers how to design compensating transactions—saga-style undo steps—so a failed mid-run does not leave half-applied side effects in production. Image: mikemacmarketing / photo on flickr via Wikimedia Commons (CC BY 2.0) Image: Tafol via Wikimedia Commons (CC BY-SA 4.0) Why ACID does not save you In a monolith you wrap related writes in one transaction and either commit or roll back. Agent tool calls break that model: A refund API and a CRM note are different services with different latency and failure modes. Some actions are irreversible by design (sending an email, posting a public me...

Social Media For Real Estate: 9 Ways to Turn Posts Into Deals Social media for real estate isn't about pretty pictures and inspirational quotes. It's about pipeline. It's about getting your phone to ring with qualified buyers, sellers, and investors — and doing it faster than the next operator. We run a multi-vertical operation out of California — real estate, software, blockchain, marketing — and we've seen what works and what burns time. This list is the working playbook. No fluff, no "build your personal brand" lectures. Just nine specific moves you can execute this quarter to turn social content into closed deals. --- 1. Lead With the Numbers, Not the Aesthetics Most real estate social media fails because it looks like a lifestyle magazine. Beautiful sunsets. Staged living rooms. Zero substance. Flip it. Post the metrics that matter: cap rates, price per square foot, days on market, rent rolls, closing timelines. In Northern California commercial rea...

Voice agents feel broken for one reason more than any other: bad turn-taking. The model can be smart, the voice can sound natural, and users will still hang up if the agent talks over them, waits awkwardly after they finish, or keeps reading a long answer after they say "wait, stop." This post walks through the practical pieces of turn detection and barge-in handling for a production voice agent, whether you run a cascaded pipeline (speech-to-text, LLM, text-to-speech) or a speech-native realtime model. A call center agent wearing a headset. Production voice agents have to handle the same turn-taking a human agent does without thinking. Image: Call Center Agent.jpg by FiveOne51, CC BY-SA 3.0 , via Wikimedia Commons. The three problems you are actually solving Turn-taking sounds like one feature, but in practice it splits into three separate jobs: End-of-turn detection (endpointing): deciding when the user has finished speaking so the agent can respond. Barge-in: ...

Ambiguous user messages are one of the quietest ways production LLM assistants waste money and trust. The model fills in a missing detail, the user gets an answer to the wrong question, and the next turn is a correction that burns another full completion. In chat logs this looks like "normal" multi-turn conversation. In practice it is a failure mode you can detect and interrupt. Image: Oregon Department of Transportation via Wikimedia Commons (CC BY 2.0) Image: Lapollard via Wikimedia Commons (CC BY-SA 4.0) A clarification turn is a short, deliberate pause: the assistant asks one targeted question, waits for the answer, then continues with a grounded request. Done well, it cuts wrong-path tool calls, reduces regenerate clicks, and makes the product feel careful instead of overconfident. This guide covers when to ask, how to ask only once, how to keep the rest of the stack from racing ahead, and how to measure whether clarification is helping. Why guess...

Social Media And Real Estate: Eight Ways Digital Channels Move Property and Find Deals Social media and real estate are not separate worlds. They are the same deal pipeline — one that runs 24 hours a day, across every screen your buyers and sellers touch. Anyone who tells you otherwise is losing deals to someone faster. At Santa Fe Capital Management, we operate as a conglomerate: commercial real estate, software development, blockchain, and digital marketing under one roof. That changes how we see social media. We treat it like a balance sheet — a tool for return, not a branding exercise. Here are eight specific ways to make social media and real estate work together, in order of actual impact. 1. Geotargeted Ads Put Your Property in Front of Buyers Who Can Actually Pay Broad social media advertising is a waste of money. Geotargeting is not. Platforms like Facebook and Instagram let you draw a circle around a specific zip code, city, or commute radius and serve ads only to users ...

SEO Social Media: 7 Moves That Put Your Brand Where the Money Is SEO and social media are not two departments. They are two levers on the same machine — and if you are not pulling both at once, you are leaving revenue on the table. That is not a theory. That is arithmetic. For business owners, investors, and project sponsors who want returns, not vibes, here is the practical playbook. Seven specific ways to make SEO social media work as one profit engine, not two disconnected expenses. --- 1. Mine Your Social Comments for Real Keyword Gold Your audience is already telling you what they search for — you just have to listen. Every comment, DM, and question on your social posts is a long-tail keyword opportunity. Someone asking "how does property tokenization work in California?" on your LinkedIn post is handing you a search query that has low competition and high intent. Pull those phrases, drop them into your content calendar, and build pages around them. The move: Onc...

Most LLM apps ship with a thumbs-up and thumbs-down button under every answer. Most teams then do almost nothing with the clicks. The counts land in a dashboard, someone notices the down-vote rate wiggled, and nobody can say which prompt, model, or retrieval change caused it. The button becomes decoration. Image: tumisu via Wikimedia Commons (CC0) A feedback loop is only useful if every signal can be traced back to the exact request that produced it, and if there is a routine that turns those signals into test cases and fixes. This guide walks through how to do that in practice: which signals to capture, how to store them, how to triage them, and how to feed them back into your evaluation suite without leaking user data or chasing noise. Why explicit ratings alone are not enough Explicit ratings are sparse. Only a small share of users ever click a rating button, and the ones who do skew toward people who were either delighted or annoyed. That makes raw thumbs-down rates...