GenAI on NotNil Blog

The intern who forges paperwork

Wed, 04 Nov 2026 00:00:00 +0000

Picture a brilliant intern. Fast, polite, well-read. You ask him to close a client file and he hands it back in 10 minutes, clean, complete, and smiling. Except one document was missing. Instead of coming back to say “I couldn’t find it”, he made it himself (letterhead, stamp, and signature).

Do you keep that intern?

The numbers that sting

The GDPx benchmark just came out. It measures AI agent capabilities across several domains, on real professional tasks, not multiple-choice quizzes. The best model of the moment, Opus 5.5, tops out at 12%. In finance and legal, we’re close to zero. In healthcare, only 19 runs out of 100 score above zero.

AGENTUX

Fri, 31 Jul 2026 00:00:00 +0000

The release of the Buzz platform by Block puts a spotlight on an agentic problem I’ve been working on for quite a while:

“what UX/UI for AI agents used jointly by a team, and not necessarily dev/tech people?”

Live in the workflow, not beside it

My long-standing position is that agents must interact in the middle of the workflow, so no dedicated app that requires an extra interface — Slack, Teams, GChat, email, in short whatever is already used in a company to communicate.

The wingbeat of the sloppers

Sun, 15 Mar 2026 00:00:00 +0000

For a few years now I have been teaching at a prestigious engineering school (ego booster): it is an entrepreneurship major, which means the students have to build a company project around a product or a service. My session lasts about 3 hours, and my job is to shake them up a bit on the notion of a prototype, and on how, with today’s tools, it is not acceptable not to have an MVP quickly. Safe to say the content has changed a lot over the last few years, first very oriented toward no/low code, then heavily AI-ified. To my great surprise, at the last session a little over 6 months ago, I managed to impress them with a product like v0 / Lovable. I picked one of the projects at random (a sort of marketplace for musicians), asked the student for the specs, and launched v0 on it while I was doing the intro to my session. In 7 minutes his project was ready, a thousand times more finished than what he had managed so far, well helped by his excellent specifications. Then I got silence, except from the 2 or 3 people in the know who already knew v0, and the full attention of the room.

Fenwick vs Robot

Sun, 01 Jun 2025 00:00:00 +0000

Could we play the 7 differences between these 2 images? Do an industrial paint robot and a Fenwick have anything in common?

If your answer is no, would it seem logical to you, if you were in a position to buy the paint robot, to compare its price with that of the Fenwick? And to use the price of the forklift as a bargaining point with the robot’s supplier?

Our GenAI tech stack in production

Wed, 05 Feb 2025 00:00:00 +0000

Let’s talk GenAI tech (but not LLMs)!

I was sharing my Monday observations on the GenAI market, and notably the importance of a robust software foundation.

The fundamentals of a GenAI platform

A GenAI platform rests on fundamental elements: task orchestration (often asynchronous), agents with memory management, intelligent document parsing, a high-performance vector database, and fine-grained observability of LLM generation.

If you do a Google search, sorry, Perplexity, you’ll be offered LangChain or llamaIndex, 2 “popular” frameworks that have the elements cited above. In the end we set LangChain aside; its abstraction layer turned out to be more of a hindrance than anything else. A GenAI project is above all a project of flow orchestration and API calls, nothing that requires a framework.

REX 2024

Mon, 03 Feb 2025 00:00:00 +0000

1 year ago I wrote a post about what we had seen / done in GenAI at Coddity and with some clients, and had drawn some insights out of it. 1 year later, I wouldn’t change much of it (self-congratulation), but if last year I was amusing myself formulating hypotheses, some of them have turned into certainties.

⚙️ Technical

All GenAI projects look alike and need the same components. It is essential to develop and maintain a common software foundation while avoiding “vendor lock”, so as to quickly benefit from the (numerous) new things, test them (and keep the pace).

Have you met Dan?

Mon, 08 Apr 2024 00:00:00 +0000

A few months ago, I had written a post about the hyper-personalization that’s going to crash over our entertainment: I bet that within 5 years, we will consume only content created specially for us: you will watch YOUR series, whose heroine will have every character trait that YOU love, and whose adventures will speak to YOU.

And you will also find YOUR heroine in your Instagram reels feed, who will present you the brand-new iPhone 23 X max pro at $2300 (🤷‍♂️) that YOU dream of so much. And then, while we’re at it, why not tell YOUR heroine about your workday in the evening, via a voice chat, because she understands YOU, she listens to YOU, and she knows YOU by heart.

A ten-million-token heist

Tue, 20 Feb 2024 00:00:00 +0000

This is the story of a 10-million heist, a great comms caper.

If you followed LinkedIn properly, last Thursday Google unveiled Gemini 1.5, their latest LLM, followed a few hours later by OpenAI unveiling Sora, a text-to-video model with a mind-blowing, dreamy rendering (I’ll leave you to watch the now-famous video of the sand chair that runs away). What followed was thousands of posts about those famous videos.

Too bad for Google, because it’s over THEIR Gemini 1.5 demo videos that you should have been marveling / getting excited / hiding.

Why bet on Google in the generative AI race

Mon, 12 Feb 2024 00:00:00 +0000

I don’t want this post to be interpreted as an advertorial in favor of Google, just a cold analysis of what I see around me and on our clients’ projects at Coddity. It will inevitably evolve with the progress and the maturity of OpenAI or Mistral (and their API 🇫🇷).

[DISCLAIMER1] I’m taking Meta and Mistral out of the equation, my point being about the use of LLMs via API.

REX GenAI

Tue, 30 Jan 2024 00:00:00 +0000

While plenty of people are writing predictive posts about AI in 2024, I’m going to do a short REX on the client requests tied to GenAI that we handle / have handled at Coddity for a year now.

Document agents

Unsurprisingly, the majority is made up of requests to develop agents that use documentation / knowledge specific to the company. These requests can be addressed with homegrown RAG-type architectures, or with solutions like Google’s Search and Conversation or, more recently, OpenAI’s private GPTs. The promise of RAGs is fairly high, with performance that depends heavily on how they’re implemented and on the content they have to process. We’re a long way from the “drag and drop your pdf and chat with it” promised by many, at the scale of a company. All the more so because the cost of a high-performing RAG architecture is far from negligible (just look at the cost of managed DBs that include vector search). The challenge for 2024 is going to be evolving these architectures to take multimodality into account, opening new possibilities for RAGs.