Filipp Trigub
@filipptrigub.bsky.social
Building the next AI-native YouTube on http://huuh.me
AI engineer, physicist, futurist, activist, hobby philosopher, infrequent ranter
Sharing my work and occasional insights.
AI engineer, physicist, futurist, activist, hobby philosopher, infrequent ranter
Sharing my work and occasional insights.
Communities can be a solution, but only if they are the perfect size.
Yes, the feed is the money machine, but even time range filters or the ability to exclude "person x liked this" posts would make the experience more enjoyable.
Yes, the feed is the money machine, but even time range filters or the ability to exclude "person x liked this" posts would make the experience more enjoyable.
August 16, 2025 at 7:23 AM
Communities can be a solution, but only if they are the perfect size.
Yes, the feed is the money machine, but even time range filters or the ability to exclude "person x liked this" posts would make the experience more enjoyable.
Yes, the feed is the money machine, but even time range filters or the ability to exclude "person x liked this" posts would make the experience more enjoyable.
The recent feed includes so much that it's borderline useless.
If there even is one on the given platform.
If there even is one on the given platform.
August 16, 2025 at 7:23 AM
The recent feed includes so much that it's borderline useless.
If there even is one on the given platform.
If there even is one on the given platform.
The popular posts on the main feed are usually days or weeks old.
Commenting rarely creates a conversation.
The only thing that happens is you getting addicted to the feed and feeling bad about it.
Commenting rarely creates a conversation.
The only thing that happens is you getting addicted to the feed and feeling bad about it.
August 16, 2025 at 7:23 AM
The popular posts on the main feed are usually days or weeks old.
Commenting rarely creates a conversation.
The only thing that happens is you getting addicted to the feed and feeling bad about it.
Commenting rarely creates a conversation.
The only thing that happens is you getting addicted to the feed and feeling bad about it.
ok, you carry on then :)
August 13, 2025 at 6:19 PM
ok, you carry on then :)
notebooklm is by google...
August 13, 2025 at 6:04 PM
notebooklm is by google...
You can also see that search-enhanced tools clearly performed much better.
You can further enhance that via various techniques like RAG, CAG or other data integrations.
Many tools that can do that: notebooklm, huuh(.)me, etc
You can further enhance that via various techniques like RAG, CAG or other data integrations.
Many tools that can do that: notebooklm, huuh(.)me, etc
August 13, 2025 at 5:40 PM
You can also see that search-enhanced tools clearly performed much better.
You can further enhance that via various techniques like RAG, CAG or other data integrations.
Many tools that can do that: notebooklm, huuh(.)me, etc
You can further enhance that via various techniques like RAG, CAG or other data integrations.
Many tools that can do that: notebooklm, huuh(.)me, etc
Given the decentralized, manifold and at times unreliable nature of climate tech, we will need a lot of predictive AI to help stabilize that.
August 13, 2025 at 5:30 PM
Given the decentralized, manifold and at times unreliable nature of climate tech, we will need a lot of predictive AI to help stabilize that.
Wait...tiktok has moderators?
August 13, 2025 at 5:28 PM
Wait...tiktok has moderators?
This is so 2023.
Yes, stochastic models are stochastic.
But give them the right context and they will actually provide a sufficiently good answer.
People need to upgrade from chatGPT to better tools.
Yes, stochastic models are stochastic.
But give them the right context and they will actually provide a sufficiently good answer.
People need to upgrade from chatGPT to better tools.
August 13, 2025 at 5:27 PM
This is so 2023.
Yes, stochastic models are stochastic.
But give them the right context and they will actually provide a sufficiently good answer.
People need to upgrade from chatGPT to better tools.
Yes, stochastic models are stochastic.
But give them the right context and they will actually provide a sufficiently good answer.
People need to upgrade from chatGPT to better tools.
How about we solve the measurement problem first...
August 13, 2025 at 5:24 PM
How about we solve the measurement problem first...
Assigning founders (or investors for that matter) archetypes is heuristic that only works on average, thus cannot be invoked to argue about individual cases. Use it to analyse 1000 founders, OK, but don't assume anything about individual founders based on it.
August 2, 2025 at 8:40 PM
Assigning founders (or investors for that matter) archetypes is heuristic that only works on average, thus cannot be invoked to argue about individual cases. Use it to analyse 1000 founders, OK, but don't assume anything about individual founders based on it.