Truecaller faces mounting pressures as its growth matures
As growth slows, Truecaller is leaning on subscriptions, business services, and new features to sustain momentum beyond India.
TechBrief — تازهترین اخبار فناوری
آخرین خبرها
The Prompt API
Article URL: https://developer.chrome.com/docs/ai/prompt-api
Comments URL: https://news.ycombinator.com/item?id=47917026
Points: 54
# Comments: 36
EvanFlow – A TDD driven feedback loop for Claude Code
Article URL: https://github.com/evanklem/evanflow
Comments URL: https://news.ycombinator.com/item?id=47916909
Points: 42
# Comments: 16
TurboQuant: A first-principles walkthrough
Article URL: https://arkaung.github.io/interactive-turboquant/
Comments URL: https://news.ycombinator.com/item?id=47916890
Points: 71
# Comments: 10
Google banks on AI edge to catch up to cloud rivals Amazon and Microsoft
Article URL: https://www.ft.com/content/2429f0f0-b685-4747-b425-bf8001a2e94c
Comments URL: https://news.ycombinator.com/item?id=47916410
Points: 89
# Comments: 60
FreeBSD Device Drivers Book
Article URL: https://github.com/ebrandi/FDD-book
Comments URL: https://news.ycombinator.com/item?id=47915632
Points: 57
# Comments: 7
The Stanford freshmen who want to rule the world . . . will probably read this book and try even harder
Can a book like this actually change anything? Or does the spotlight, as it always seems to, send more students racing to the place?
Amazon’s new podcast strategy: Monetize everything
Amazon's podcasting business seems to have transformed over the past six months.
Butterflies are in decline across North America, a look at the Western Monarch
Comments URL: https://news.ycombinator.com/item?id=47914677
Points: 182
# Comments: 55
Show HN: AI memory with biological decay (52% recall)
Most RAG setups fail because they treat memory like a static filing cabinet. When every transient bug fix or abandoned rule is stored forever, the context window eventually chokes on noise, spiking token costs and degrading the agent's reasoning.
This implementation experiments with a biological approach by using the Ebbinghaus forgetting curve to manage context as a living substrate. Memories are assigned a "strength" score where each recall reinforces the data and flattens its decay curve (spaced repetition), while unused data eventually hits a threshold and is pruned.
To solve the "logical neighbor" problem where semantic search misses relevant but non-similar nodes, a graph layer is layered over the vector store. Benchmarked against the LoCoMo dataset, this reached 52% Recall@5, nearly double the accuracy of stateless vector stores, while cutting token waste by roughly 84%.
Built as a local first MCP server using DuckDB, the hypothesis is that for agents handling long-running projects, "what to forget" is just as critical as "what to remember." I'd be interested to hear if others are exploring non-linear decay or similar biological constraints for context management.
GitHub: https://github.com/sachitrafa/cognitive-ai-memory
Comments URL: https://news.ycombinator.com/item?id=47914367
Points: 81
# Comments: 38