Featured2 min read
How we cut LLM inference costs by 61% without touching model quality
Prompt caching, request batching and a routing layer did more than any model swap. Here is what we measured, in order.
0 articles7 subtopics
Topic
Technology
How software, infrastructure and data actually get built and run.
Most read
Featured1 min read
Why we still keep a human in the loop for refunds over $200
Our agent could technically approve these on its own. The data on why we still do not is more interesting than the policy.
Featured1 min read
What a tabletop incident response exercise actually looks like
A step-by-step walkthrough of the two-hour exercise we run with clients, including the scenario we use and where teams usually get stuck.
Technical deep dive1 min read
The support ticket router that took us three tries to get right
Our first two attempts at automatic ticket routing made things worse. The third one shipped because we stopped optimizing for accuracy.
Latest
Guide7 min read
The Ultimate Guide to Discord Bots
Discord bots have revolutionized the way communities interact and engage with each other on the popular communication platform, Discord.
Opinion3 min read
The Job Search Has Changed. And We're Optimizing for the Wrong Thing
You can have 10+ years in your field, a strong CV, the right certifications, and a track record of actually delivering results and still never get a chance to speak to a human.
Guide4 min read
How to build a workflow you can actually inspect
A walkthrough of building your first automation in Nodesin, and why every step, its inputs and outputs, and the bill remain visible.
Featured1 min read
Detecting phishing before it reaches your employees
Most phishing that gets through is not sophisticated. It is fast. A practical layered setup for a 50-person company.
Guide1 min read
The five email security controls that stop 90% of incidents
Ranked by the incidents we have actually responded to, not by how they are usually marketed.
Opinion1 min read
What we tell customers before turning on an AI agent
Six sentences we make every customer read and confirm before we enable autonomous replies. Most of the value is in what we exclude.
Featured1 min read
Postgres full-text search is enough until it isn't. Here is where the line is.
We put off adding a dedicated search service for two years. A tutorial on getting real search out of Postgres, and the signals that told us to move on.
Analysis1 min read
Why your MFA rollout probably has a gap
We have audited over 60 companies that believed MFA was fully enforced. Fewer than ten actually were.
Guide1 min read
Evaluating LLM output when there is no ground truth
Most of our extraction tasks have no single correct answer to grade against. This is the scoring approach we landed on after two that failed.