Du Paris kabyle aux Kabyles de Paris, conférence

https://fed.brid.gy/r/https://ahavparis.com/paris-kabyle-aux-kabyles-de-paris-conference/

Have you (or your company) ever built something different in tech, something that could make the world even 1% better, but struggled to talk about it?

Maybe the idea felt too big, or too early.

Maybe you weren’t sure how to shape it into a conference talk or a podcast conversation.

You’re not the only one. And your work deserves to be heard.

There’s a stage waiting for ideas like yours: https://TechDifferently.com

#TechDifferently #PublicSpeaking #technology #conferences #startups

Title: P4: I was at the biggest IT party-conference in Moscow. 🤖 [2025-08-16 Sat]

- Cognitive load – lower is better
- Flow state – higher is better
- Feedback loops – earlier is better

One speaker sees the AI future in: 🚀
- Edge computing
- AI assistants-coworkers
- Evolutionary, adaptive applications
█ #dailyreport #conferences #synergy #learning #llm #agents #reasoning

Title: P3: I was at the biggest IT party-conference in Moscow. 🤖 [2025-08-16 Sat]

Learning learns by copying experts, often used to "warm
start" RL, but sometimes combined for better
performance. This is simular to how LLM trained first by
self-supervised learning to predict next token and then
by RLHF.

Interesting DevEx practice 👨‍💻 shifts focus from
measuring development output to how the developer feels: #dailyreport #conferences #synergy #learning #llm #agents #reasoning

Title: P2: I was at the biggest IT party-conference in Moscow. 🤖 [2025-08-16 Sat]

cells ꙮ act as small soldiers able to adapt, and how
flexible biological programs may be.

In the self-driving car 🚗 speech, I noted an interesting
fact: agents have “prediction” to imagine the near
future before the reasoning step.

RL learns by trial and error with rewards, Imitational #dailyreport #conferences #synergy #learning #llm #agents #reasoning

Title: P1: I was at the biggest IT party-conference in Moscow. 🤖 [2025-08-16 Sat]

AIRI suggests ways to increase AI agent robustness:
1) Full monitoring (with recording current state in plan)
2) Replanning: add rules to avoid doing something bad,
fallback strategies

The most challenging topic is to make the AI agent
improve its own code.

There was a demonstration of how biological embryonic #dailyreport #conferences #synergy #learning #llm #agents #reasoning

Title: P0: I was at the biggest IT party-conference in Moscow. 🤖 [2025-08-16 Sat]

The AIRI Institute head mentioned a very interesting
idea: that LLM reasoning is a form of task
decomposition. It explains why LLMs require additional
enforcement.
I think reasoning is not one process but three:
1) Chaotic search in memory
2) Decomposition heuristic that includes 1. as a substep
3) Logic-checking heuristic #dailyreport #conferences #synergy #learning #llm #agents #reasoning