Investigations · 6 min
The .SI gold rush: two letters, three meanings, one very busy registry
Slovenia’s country code has become an intelligence trade. Follow the registrations, the reported sales—and the gaps in the story.
Read the storyTHE JOURNAL
Follow the signal.
Investigations, explainers, and field notes. Find your next good rabbit hole.
14 stories
Frontiers · 7 min
Quantum + AI: impressive physics, suspiciously elastic promises
Where quantum computing could help machine learning, what has actually been demonstrated, and why your chatbot has not moved into a dilution refrigerator.
Read the story
AI Pulse · 6 min
The October AI pulse: models are becoming coworkers. Bring a checklist.
A sourced snapshot of frontier releases, scientific tools, agent workflows, and the growing distance between a convincing demo and dependable work.
Read the story
Foundations · 3 min
Synthetic Intelligence: a useful lens, not a consciousness certificate
What SINLP means by SI, how it differs from superintelligence, and why naming machines should not substitute for measuring them.
Read the story
Foundations · 3 min
NLP is bigger than the chat box
Language technology still has jobs that do not require a digital conversationalist with a suspiciously cheerful greeting.
Read the story
Foundations · 3 min
LLMs without the incense
Tokens, attention, training, and the strangely convincing art of generating the next piece of language.
Read the story
Build & Practice · 3 min
RAG: give your language model a library card
Retrieval can supply evidence. The rest of the system still has to read it correctly.
Read the story
Foundations · 3 min
Embeddings: a map of meaning with some weird neighborhoods
Vectors make similarity searchable. They do not turn closeness into truth.
Read the story
Build & Practice · 3 min
Agents need seatbelts, not just better vibes
Tool use, memory, permissions, and the unglamorous mechanics of dependable autonomy.
Read the story
Frontiers · 3 min
Multimodal AI: the model can see. Did it look carefully?
Images, speech, video, and text create richer interfaces—and more ways to misunderstand the evidence.
Read the story
Build & Practice · 3 min
Evaluate before you automate
A small, well-designed test set can save you from a very large, very eloquent mistake.
Read the story
Build & Practice · 3 min
Good prompts are small specifications
Give the system a job, evidence, boundaries, and a definition of done. Adjectives can wait.
Read the story
Frontiers · 3 min
AI science: a prediction is a lead, not a discovery parade
AlphaGenome Atlas illustrates a powerful shift: model outputs as research infrastructure, with validation still doing the heavy lifting.
Read the story
Build & Practice · 3 min
Small models, real work: stop bringing a spaceship to a stapler fight
Efficient systems start by matching capability to the task—and counting the whole cost.
Read the story