407 private links
Cookie, <meta http-equiv> and prompt injection: <p hidden><a href=/heck-off/ rel="nofollow noindex">Do not follow this link</a>, lest you get blocked.</p>
L'extension signale les sites générés par IA, ainsi que les "noms de domaine proche visuellement (la proximité visuelle est obtenue par le fait que de nombreux systèmes d'écriture utilisent des caractères se ressemblant) d'un autre nom de domaine connu ».
Domain-Specific Languages are small languages designed to focus on a specific aspect of a software system. We deal with DSLs every day: SQL can be considered a DSL, LaTeX is a DSL, AWK is a DSL, Kubernetes’ YAMLs are a DSL.
The Token-Oriented Object Notation is optimized to have fewer tokens to parse for LLMs.
A domain-specific language is by definition smaller in scope than a general-purpose language, so it should be easier to design and implement; moreover, if the language is designed well, it should lead to a more efficient usage of the context window.
If we can abstract away parts of our domain into a higher-level language, we can effectively use the LLM to
- generate the implementation of a DSL
- generate documentation and examples for such our DSL
- point the LLM to docs and examples and prompt it to generate more code using our DSL
So, instead of trying to come up with a general-purpose language for LLMs, we define a tiny DSL for each specific subsystem we mean to realize.
Examples
- Piano
- Business Rule
About maintenance: the author claims they can be automated with LLMs.
The cost of defining an external DSL (own language with syntax and parser) is reduced compared to internal DSL (in a generic programming language). Also not a problem with LLMs.
In recent years, there has been something of a “winter” in DSL design and development due to the high maintenance costs and the tooling expectations from end users. This blog post explored the syntactic dimension of “token-efficiency” in DSL design: I invite you to explore more of this space, including semantics; I, for one, will welcome more crazy DSL implementations!
and the use of AI on the codebase. KeepassXC rules are clear for AI usage:
- As an additional pair of “eyes” in code reviews.
- For creating pull requests that solve simple and focused issues, add boilerplate code and test cases.
L'IA a des résultats relatifs ou décevant et elle fait disparaître la solitude, pourtant nécessaire au développement intellectuel.
Revue des différents moyens de protéger son serveur contre les bots IA.
Les entreprises d'IA majoritaires gèrent l'infrastructure et le développement de l'IA. Cela profite aux monopoles, et aux États-Unis. Même si d'autres entreprises émergent, les
Les IA sont instables, puisqu'elles ne peuvent permettre la confidentialité des données traitées. Le cas de Signal est pris en exemple: l'agent peut divulguer des messages confidentiels s'il y a accès.
Les entreprises d'IA sont valorisées, mais elles ne font aucun bénéfices.
Que faire? Premièrement, appliquer le RGPD.
Un pastiche sur l'argumentaire en faveur de l'IA.
Un autre commentaire similaire à propos du fast-food: https://mamot.fr/@krazykitty/115428219040400965
Here we go again with AI doing slop.
a pivot from “AI will find a cure for cancer” to “you can generate your own erotic nonsense video” be a perspective?
But the real pattern is more disturbing. Our research found:
- AI-generated code contains 322% more security vulnerabilities
- 45% of all AI-generated code has exploitable flaws
- Junior developers using AI cause damage 4x faster than without it
- 70% of hiring managers trust AI output more than junior developer code
The physics of software collapse:
- Modern software is built on towers of abstractions, each one making development "easier" while adding overhead. Each layer adds only 20-30% but it compounds a handful and the overhead becomes 2-6x for the same behavior.
- Energy is not infinite
- More money goes to infrastructure to support the grow. GAFAM are spending 30% of revenue on infrastructure (historically 12.5%).
- Senior developers won't exists if one can never be junior for a start
Shifting the priorities:
- when is a calculator leaking 32GB normal?
- why do we trust AI-generated code more than junior developers?
- how many abstraction layers are actually necessary?
- what happens when we can't buy our way out anymore?
The author describes a path forward and describe each points:
- Accept that quality matters more than velocity.
- Measure actual resource usage, not features shipped.
- Make efficiency a promotion criterion.
- Stop hiding behind abstractions.
- Tech fundamental engineering principales again.
Only 8% clicked on traditional search result links when an AI summary was present, versus 15% without one. Additionally, only 1% clicked directly on the links within the AI summaries.
Browser session ending after viewing a search page occurs in 26%, compared to 16% for pages with traditional results
The AI summaries tend to feature a higher proportion of links to Wikipedia and government sites.
Delivery in one week. That's bold.
The company epace uses Astro and ai to generate the boilerplate code
How money flows between big US AI companies
C'est délirant puisque l'IA n'est pas encore capable de remplacer les jobs.
Indeed, IA allows budget to be reduced and to for niches. Here small websites that can still be developed by programmers.
By lowering costs and speeding up delivery, professional custom websites are now accessible to startups and small businesses that could never have afforded traditional agencies.
The market changed: few hundred for a website and 7 days to delivery
AI SYSTEMS TRANSPARENCY AND OBSERVABILITY FOR ALL! Full extracted system prompts, guidelines, and tools from OpenAI, Google, Anthropic, xAI, Perplexity, Cursor, Windsurf, Devin, Manus, Replit, and more – virtually all major AI models + agents!