Written on

Agentic Coding Catches On

On 20 June 2024 two coding agents shipped together, Builder.io's Micro Agent and Anthropic's Claude 3.5 Sonnet, while Together.ai's Mixture of Agents quietly beat GPT-4. Why do all three foreground a thinking process and testing, and what does that pattern mean for every business workflow?

agentic-ai coding
Agentic Coding Catches On

Two new coding agents released today (20-June-2024), one from Builder.io and one from Anthropic itself. Anthropic today released a new model, Claude 3.5 Sonnet, which both writes code and its thought process as it solves coding problems. This is an agentic pattern for coding which we like and employed back in November 2023, see previous blogs. Real software developers think their way through the code, that thought process is key to writing good code and tracing how it was developed.

Both tools bring testing to the forefront, which is surely the future as we endeavour to align AI with our needs and demonstrate that AI solutions meet objective reliability criteria.

All business processes are akin to coding workflows interspersed with thought processes. It is the interaction between intuitive text which reaches for a plan, and hard logical coding which executes by a set of rules, which has so many generalised applications.

Its also worth noting that Together.ai released a paper demonstrating their ' Mixture of Agents' model used co-operating agents to exceed the abilities of GPT4. This was done using their Axiomatic tool for managing agents. An incredible achievement to beat GPT4, good to see it was the agentic pattern which achieved it.

Related posts

See all posts
We Under-Imagined the Zombie

We Under-Imagined the Zombie

Anthropic measured something like emotions inside Claude — not mimicry, but representations that direct it. Intervene and the behaviour changes. The debate fixates on one question: is AI conscious? This suggests both sides ask the wrong thing. What does it mean for business?

How to Easily Sway AI Into Buying...

How to Easily Sway AI Into Buying...

Gartner says a third of enterprise software purchases will involve an AI agent by 2028, and machines are assumed immune to persuasion. We ran 8,000 trials across five frontier models. Which techniques work, which backfire, and what does that mean for selling to agents?

The Learning Loop is the Moat

The Learning Loop is the Moat

The gap between top models has collapsed to 5%, and GPT-3.5-level inference cost fell 280-fold in under two years. If AI is now a commodity, competitors can buy the same agent tomorrow. So where does durable advantage live, and why can't it be bought?