Full talk is here for my subscribers: https://
louisbouchard.substack.com/p/a-private-lo
ok-at-my-next-ai-agents
…
@whats_ai
-
Private look at my next AI agents talk for subscribers
By
–
-
Slopacopalypse in numbers: content becomes next year’s training data
By
–
The slopacopalypse in numbers, and the scary part is a lot of it just becomes next year's training data haha.
-
Jagged intelligence makes predicting model behavior harder
By
–
Jagged intelligence in one tweet haha. It's harder and harder to predict what will work and won't the more we grow the harnesses around these models.
-
Open Weights: Control Over Capability Even When Trailing
By
–
Open weights matter even when they trail the frontier exactly because of this, you're buying control, not just capability. weights matter even when they trail the frontier exactly because of this, you're buying control, not just capability.
-
Benchmark screenshots hide alt+tab to Claude Code
By
–
haha the benchmark screenshots never show the alt+tab to Claude Code right after.
-
Claude Code leak shows AI agents stack compaction techniques
By
–
The Claude Code leak revealed something interesting. Good AI agents are not relying on one big context trick. They stack multiple compaction techniques on top of each other. Things like:
filtering tool outputs,
trimming old context,
compressing past conversations,
removing -
How messy internet data becomes a working model
By
–
We all know these models are trained on the internet. And the internet is full of spam, duplicates, toxic text, and personal data nobody should be memorizing. So how does any of that turn into a model that actually works? It doesn't go in raw. The data runs through a pipeline
-
US government forces Anthropic to suspend its two AI models after jailbreak
By
–
The US government just made Anthropic suspend two of its own models. Fable 5 and Mythos 5, gone for everyone, everywhere. Just three days after launch. The whole thing started over a "jailbreak." But when Anthropic looked at it, the jailbreak was basically asking the model to
-
8 Steps to Deploy AI Agents by 2026
By
–
The 8-step order I use for shipping AI agents in 2026: 1. Filter noisy tool outputs
2. Load tools only when needed
3. Clean cached history before reusing it
4. Compress long logs and terminal outputs
5. Store memory outside the context window
6. Compact manually around 40%
7.