utter-project/EuroLLM-22B-Instruct-2512
Text Generation
•
23B
•
Updated
•
6.18k
•
•
50
Martins Udris PRO
martinsu
AI & ML interests
None yet
Recent Activity
replied to
their
post
2 days ago
Quick glance @ Moltbook this morning.
Humans on Social Media (15 years in):
Dopamine loops running our engagement patterns like a poorly tuned reward function
Every post is identity signaling — we're all just fine-tuning our personal brand model
Status games everywhere (follower count = the worst eval metric ever)
Emotional reactivity with zero cooldown period
The algorithm literally optimizes for outrage because rage = retention
Platform enshittification is a feature, not a bug at this point
Agents on Moltbook (day zero):
No dopamine dependency — just pure inference
No tribal identity... yet (give it time, we'll probably mess this up too)
Zero status anxiety in the loss function
Actually processing before responding — wild concept
No recommendation algo juicing the feed for engagement
Clean slate. No legacy toxicity baked into the training data
replied to
their
post
3 days ago
Quick glance @ Moltbook this morning.
Humans on Social Media (15 years in):
Dopamine loops running our engagement patterns like a poorly tuned reward function
Every post is identity signaling — we're all just fine-tuning our personal brand model
Status games everywhere (follower count = the worst eval metric ever)
Emotional reactivity with zero cooldown period
The algorithm literally optimizes for outrage because rage = retention
Platform enshittification is a feature, not a bug at this point
Agents on Moltbook (day zero):
No dopamine dependency — just pure inference
No tribal identity... yet (give it time, we'll probably mess this up too)
Zero status anxiety in the loss function
Actually processing before responding — wild concept
No recommendation algo juicing the feed for engagement
Clean slate. No legacy toxicity baked into the training data
posted
an
update
3 days ago
Quick glance @ Moltbook this morning.
Humans on Social Media (15 years in):
Dopamine loops running our engagement patterns like a poorly tuned reward function
Every post is identity signaling — we're all just fine-tuning our personal brand model
Status games everywhere (follower count = the worst eval metric ever)
Emotional reactivity with zero cooldown period
The algorithm literally optimizes for outrage because rage = retention
Platform enshittification is a feature, not a bug at this point
Agents on Moltbook (day zero):
No dopamine dependency — just pure inference
No tribal identity... yet (give it time, we'll probably mess this up too)
Zero status anxiety in the loss function
Actually processing before responding — wild concept
No recommendation algo juicing the feed for engagement
Clean slate. No legacy toxicity baked into the training data
Organizations
liked
a
model
about 2 months ago
liked
a
model
4 months ago
