On X — we follow important research papers and explain ideas behind it. In the newsletter, we connect the dots between AI’s past, present, and future ⬇️
Join over 115,000 readers
- Tweets 20,079
- Following 8,144
- Followers 88,415
- Likes 13,701
Turing Post (Ksenia Se) retweeted
This report provides very interesting results, for example: when asked to generate 20,000 training examples instead of 2,000, Claude produced less varied data. Adaption’s dataset generator, Invent, produced more variety.
Ten times more examples doesn’t guarantee ten times more to learn from!
Nice work, Sara and team!
We released invent a dataset.
Today we release the technical report that goes with it.
Huge shoutout to everyone @singhshiviii @andrijazzz @lekeonilude @sudip_r0y 🔥
This tackles the hardest setting. How do you go from zero data regime to a training set for a capability.
4
6
17
3,056
Turing Post (Ksenia Se) retweeted
There is an interesting effect of calling your dot over the phone:
I actually want to call my parents more.
Don’t know how it works. I guess if I have time for chatting with a neural network, I can certainly have time to talk to the dearest ones?
1
1
9
3,168
Turing Post (Ksenia Se) retweeted
Have been using my Dot for about two hours, it does feel like unloading your brains.
I am even getting closer to zero inbox that happened never
Fun thing: you can ask it to create an animated avatar based on what it knows about you.
I’m an Owl, quite appropriately
2
3
9
3,004
Turing Post (Ksenia Se) retweeted
It is 100% possible today. Tesla has ‘summon’ button for years. And now the button becomes voice operated.
1
1
2
3,097
Absolutely brilliant. New genre?
Honored to announce my first LessWrong post. It's a weird one, and it may be jarring initially, but I truly believe it is worth a dedicated close-read. It is written in the form of an LLM CoT, but is entirely authored by me. Link:
lesswrong.com/posts/vzKWsEsk…
1
5
3,108
Turing Post (Ksenia Se) retweeted
New York Times suing OpenAI and lashing out on AI.
Also, New York Times
Seriously? That type of title is AI created for better SEO optimization
4
3
10
2,794
Duplicated, undocumented tools create drag and security blind spots.
@MongoDB proposes internal AI tool registries that make tools reusable and visible, and support consistent governance.
This idea is familiar from package managers - but in this case, it's a shared catalog built around a company’s own teams, data, and policies.
It's a place that gives developers the most important tracking features:
• Search for existing tools before building new ones.
Grouping tools by function helps developers and agents find what they need.
• Tool version history to help trace changes in agent behavior
• Visibility into security review status
• What tools exist and who owns them. A separate policy layer enforces access based on agent identity, team, environment, and action type.
But a registry doesn’t automatically make tools secure by itself. It gives security teams an inventory they can review and apply policies to.
So it's useful to read MongoDB’s full take on fighting tool sprawl with an internal registry: mongodb.com/company/blog/tec…
4
2
8
3,494
Turing Post (Ksenia Se) retweeted
Depending on UI, that might become one of the most useful things recently launched by an AI lab
Announcing ChatGPT Space, a collection of new collaborative experiences right in ChatGPT.
Edit rich notes with live visualizations together with others and your dots. Whether it’s a TODO list to delegate, a product plan or meeting notes. Working with others and your agents is seamless and they also be automatically kept up to date in the background. For developers, this is AGENTS.md taken one step further.
1
1
7
3,553
Connecting the dots
What DevDay 2026 reveals about OpenAI’s cloud ambitions, who gets access, and why I’m still keeping my OpenClaw turingpost.com/p/dots-openai via @Kseniase_
2
2,801
