What is LLM quantization? Simply explained.

By Simon Frey Last updated:

Running a language model with the same quality of ChatGPT within your own company requires expensive hardware.

But don’t worry, there is an optimization: Quantization can drastically reduce the hardware requirements of LLMs (up to 80%). In this post, I’ll explain how it works.

Posted in AI

Hiding a tank in plain sight. 3 attack vectors on AI systems and how they can be countered

By Simon Frey Last updated:

As you integrate AI systems into critical business processes, they become a stronger target for cyberattacks. Attackers continuously develop new methods to exploit vulnerabilities in AI models, making the security of these systems a challenge.…

Read more of Hiding a tank in plain sight. 3 attack vectors on AI systems and how they can be countered

Posted in AI

Kids don’t walk

By Simon Frey Last updated:

I witnessed this yesterday: my neighbor’s daughter stood transfixed by a butterfly landing on a fence post, her body so still she seemed part of the garden. Seconds later, she was tearing across the lawn, arms spread wide, chasing nothing but her own delight. The contrast was beautiful in its purity.

How to fix the “pg: too many connections” error. Postgres and its three connection limits

By Simon Frey Last updated:

Are you running into any of the following postgres connection limit errors pq: sorry, too many clients already pg: too many connections for database “exampledatabase” pg: too many connections for role “examplerole” Yes? Great news:…

Read more of How to fix the “pg: too many connections” error. Postgres and its three connection limits