PITTI

In light of the recent controversies around Meta’s AI bot of xAI’s Ani, but also the broader debate on model sycophancy, I’m reposting this

PITTI - Article - Artificial Intelligence : what everyone can agree on Artificial Intelligence is a divisive subject that sparks numerous debates about both its potential and its limitations. However, there's one aspect everyone can agree on: current commercial practices...

www.pitti.io/articles/ai-...

Anthropic’s recent evals for the release of Sonnet 4 and Opus 4 represent a good opportunity to re-share a blogpost from last year entitled “Artificial Intelligence : what everyone can agree on”

It touched on benchmark-gaming

It’s helpful when you know exactly what you need… and when what you need is reasonably simple … and you know how to fix the buttons that do not work and the design.
I may use the as a starting point for something else actually

It feels more like early Replit than Cursor

I gave it another chance where I know that React Typescript could be a good option (and I’ve kind of worked on this already so I can judge the choices made)

I suppose that the wow effect is there for anyone who does not know how to code… But for now it is not a productivity tool, it’s a POC

you can deploy to cloud run (not tested) so I think that Google’s strategy is clear. They will invest in it. Let’s wait for the next upgrade

So you can’t actually work on YOUR project with your files. This just builds REACT apps from scratch for you. That’s it. I tried to ask for a Django app to check, I got another tsx app

UX is a bit weird (like closing right panel on the left… many things like this). I never found how to save a project created from scratch

Projects come with specific files (even for empty templates), you can’t delete them

It would literally build an app based on a prompt. Create files, edit them, reorganize them… very cursor like experience at a shallow level.

I’ve been testing the new Google AI Studio feature “Build” as a promising mitigant to Gemini’s code slop tendency (which the AI Studio UI makes a horrible experience)

Unfortunately this is very far from ready to do actual work

Strongly advise to wait a couple of weeks (details below)

If I believe the marketing materials, the latest Claude models can work continuously for longer than the Twitter backend does

So jack morris is in fact john morris

PITTI - Blog - May 6th, 2025 While general AI hype cools, voice tech surges, often from smaller players. At the other end of the spectrum, big players focus on the critical next phase: the maturing of the product landscape, with ...

Earlier this month, I wrote about the Cambrian explosion in the voice modality (TTS and STS) and the evolution of frontier labs’ commercial strategies

www.pitti.io/blogs/not-to...

PITTI - Blog - Part Three : From Invisible Strings to Unbreakable Chains Beyond the technical infrastructure of AI lies a more profound threat: as we willingly surrender our cognitive functions to algorithms and replace genuine human connection with digital simulations, we...

Part Three tackles the conceptual dangers arising from the spread of sophisticated influence tools as AI continues to advance.

www.pitti.io/blogs/aligne...

Part Two sheds a technical light on the profound transformations in the information value chain over the past 25 years.

www.pitti.io/blogs/aligne...

PITTI - Blog - Part One : Progress in Perspective As we navigate the early days of AI, Saint-Exupéry's reflections on aviation a century ago offer an important perspective: we are still young barbarians marveling at our new toys, racing to fly higher...

Part One of this series explores the potential disconnect between technological and societal innovation, and what makes a tool truly transformative for both.

www.pitti.io/blogs/aligne...

PITTI - Blog - Aligned : Introduction After multiple stalled attempts since 2022, a recent project finally pushed this series over the line. This introduction provides the necessary background to understand my criticism of AI in this seri...

One of my mutuals here suggested a Bluesky Wednesday as a way to revive the interest for this platform.

Reposting here a blog series bringing together various thoughts and fragments posted elsewhere around AI alignment.

Some context first

www.pitti.io/blogs/aligne...

I’ve not found this platform very welcoming last time but I’m willing to give it another try.

PITTI - Article - Existential Economic Questions for Generations X, Y and Z Generations X,Y and Z are faced with a number of existential economic questions for the rest of the century. From the looming demographic crisis in developed nations to the precarious state of soverei...

Reposting here an essay from early Feb on existential economic questions for generations X, Y and Z.

Touching on:
- Aging population
- Sovereign debt
- Technology
- A new world order

www.pitti.io/articles/exi...

PITTI - Article - Artificial Intelligence : what everyone can agree on Artificial Intelligence is a divisive subject that sparks numerous debates about both its potential and its limitations. However, there's one aspect everyone can agree on: current commercial practices...

This place is clearly not a safe haven as its decentralized nature is a double-edged sword

AI seems even more divisive than on other platforms. No punches pulled over here (and it’s not great)

In September, I wrote about what everyone can agree on.

www.pitti.io/articles/ai-...

Thanks for point this out. I’m running it locally in 8bit (in MLX), in theory it should be better than in the blog.

It is somewhat helpful but it makes so many small mistakes that it actually takes longer to fix than to write it from scratch with a bit of autocompletion

Qwen2.5-Coder-32B doesn’t really live up to expectations

Can’t handle a redux store in a fairly simple app with 3 components

When a new model comes out and I want to vibecheck it with a single prompt, my prompt is :

Please provide a detailed analysis of the strengths and weaknesses of the business model of a typical hairdresser

I feel that the situation is slightly different for commercial services where they force you to fill in a form to place an order, collect data that they don’t need and never delete them.
But I have no solution to offer. It is what it is. Same for cookies (it’s super bad but not as bad as banners)

Oh yes, I think there are a lot of lessons to be learnt from GDPR. Good intentions but completely disconnected to the practical realities (it’s even practically impossible to enforce the law). For social media platforms, Education of end-users is the only way to protect them.

lawmakers naively thought that, by putting so much burden on companies, companies will choose not to collect data and/or delete quickly

In reality they chose to put horrible banners to make users pay with their time for EU lawmakers shortsightedness and collected data like before and never deleting

And to be clear, the rationale for data laws in europe is not “don’t touch my data”, it’s “guarantee data-ownership to individuals”. The way it is implemented is a disaster. But the right ask for location, deletion and portability are good imo. Portability is underrated (and never enforced)

Presenting it like this probably misses a lot of the historical context but it’s an important factor. You understand why the Snowden revelations shocked Europe

Younger generations are more comfortable with it but they are not the ones making the law.

And what about a merge of 2 open LLMs trained on synthetic data produced by a open LLM trained on private data?

And what about the code to train the open LLM, which was itself 75% generated by a LLM?

It’s over