Maria Sukhareva — Anti-Hype AI | 35K+ | Keynotes | Workshops | AI Use Cases Reality Check All views are solely my own
Anti-Hype AI | 35K+ | Keynotes | Workshops | AI Use Cases Reality Check All views are solely my own
Maria Sukhareva ranks #204 of 18,566 LinkedIn creators in Information Technology & Services, and is a standout voice in Germany. They have 36.5K followers and published 43 posts in the last 60 days at a 0.2% average engagement rate.
- 36.5K followers
- 43 posts / 60d
- 0.2% avg engagement
- 1.3K follower growth / 30d
The roast
Maria calls herself an AI Realist, yet she managed to churn out fifty posts last month just to convince 0.24 percent of her followers that she exists. It turns out the only thing she’s been working on since before it was cool is a desperate need for external validation.
About Maria
I’m a researcher at heart, focused on applied AI. I’ve been working on language modelling since before it was cool. Check out my substack newsletter msukhareva.substack.com Disclaimer: all views posted on this account are solely my own and do not represent my employer
Highlights
- Top 1% in Information Technology & Services — Ranked #7 of 1910 creators
- Top 1% in Germany — Ranked #3 of 762 creators
- Top 5% Creator — 43 posts in 30 days
- Top 5% Audience — 36,478 followers
Recent posts
Can you imagine the panic if it were Moonshot, DeepSeek, MiniMax etc. Everyone would be like - omg, evil Chinese companies did it on purpose or they are lying - no way it happened. Mysteriously, nothing escapes from the Chinese labs. Kimi K3 doesn’t even score high on cybersecurity. (And if you are going to tell me that it’s because it was distilled from harmless Fable, read my latest article first) And I have a strong feeling that Chinese models even though having the same capabilities will not be randomly “escaping” like a little angry gnome, because AI researchers in China have heard of
129 reactions · 38 comments · 0 reposts
I hope he will register for my “Optimise Your AI Stack” workshop. But he’d better hurry up, only 5 out of 20 spots left.
15 reactions · 1 comments · 0 reposts
I have my OpenClaw running on the GLM-5.2 Coding Plan It has one job now: max out Fable 5 in Claude Code while it’s available. It’s such a rewarding experience steering Fable with GLM-5.2 straight from your Telegram chat. You can feel the future of work. And yes, for those affected, Fable did delete the event for the “Optimize Your AI Stack” workshop again. So, if you subscribed and wondered why it was cancelled for two minutes, it’s because AGI got bamboozled by Stripe webhooks and cleaned everything up. But seriously, I love this setup. Two very good models. And that’s how I use #Ope
20 reactions · 1 comments · 0 reposts
Last updated 2026-08-01