GPT-3 and the Autonomous Network

GPT-3 has convinced many that true AI is closer than we think and that it has the potential to disrupt most industries.

I wonder how this is going to impact the telecom industry…

Are we now one more step forward towards the autonomous network?

What is GPT-3?

GPT-3 is the newest AI language model from OpenAI and the most powerful language model ever built. It has 175 billion parameters, which represents over 100 times its predecessor.

The potential of this technology is unimaginable and use cases are limitless.

I am particularly interested in two main use cases that can play a fundamental a role in the autonomous network:

1. Manipulating and extracting information from text, and
2. Automating code generation.

Some practical examples

Here are some mind-blowing examples I came across these days.

All of these examples were developed by people around the globe a few days after the technology was available.

Natural Language to SQL

A GPT-3 interface to start writing SQL queries for you! Translates a natural language query into a SQL sentence – @FaraazNishtar

Semantic Search

A fully functioning search engine on top of GPT-3. For any arbitrary query, it returns the exact answer AND the corresponding URL – @parashopra

Parse Unstructured Data

A GPT-3 parser for unstructured data. Detects patterns and classifies data – OpenAI Web

Speech to Bash

Conversational bash programming interface built on GPT-3. Transforms natural language into bash code – @super3

Speech to Code

A GPT-3 interface to design a functioning app by just describing it in natural language – @sharifshameem

GPT-3 and the Autonomous Network

Today, most telecom operators are running digital transformation projects, redefining their existing culture, business processes, and customer experience.

From the network perspective, projects involve network function virtualization (NFV), simplification of network architectures, cloudification, software defined networking (SDN) and life-cycle orchestration and automation.

But getting the architecture right and automate are the very fundamentals of the autonomous network.

A white paper from TM Forum defines 6 steps to reach the Full Autonomous Network:

Conclusions

The ultimate goal of the autonomous network is the self-operation, which means:

1. Deploying Self-X capabilities (self-serving, self-fulfilling, self-assuring)
2. Providing Zero-X experience (zero-wait, zero-touch, zero-friction)

Autonomous networks need to take advantage of AI and big data to make decisions based on predictive or active closed-loop management of service and customer experience

Will GPT-3 make the dream of the autonomous network come true?

It will depend only on our capacity as an industry to articulate the use of now available technologies towards the use case.