frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Cortexa – Bloomberg terminal for agentic memory

https://cortexa.ink
7•PrateekRao01•1h ago
Hi HN — I’m Prateek Rao. My cofounders and I built Cortexa, which we describe as a Bloomberg terminal for agentic memory.

A pattern I keep seeing: when agents misbehave, most teams iterate on prompts and then “fix” it by plugging in a memory layer (vector DB + RAG). That helps sometimes — but it doesn’t guarantee correctness. In practice it often introduces a new failure mode: the agent retrieves something dubious, writes it back to memory as if it’s truth, and that mistake becomes sticky. Over time you get memory pollution, circular hallucination loops, and debugging turns into log archaeology.

What Cortexa does:

1. Agent decision forensics (end-to-end “why”): trace outputs/actions back to the exact retrievals, memory writes, and tool calls that caused them.

2. Memory write governance: intercept and score memory writes (0–1), and optionally block/quarantine ungrounded entries before they poison future runs.

3. Memory hygiene + vector store noise control: automatically detect and remove near-duplicate / low-signal entries so retrieval stays high-quality and storage + inference costs don’t creep up.

Why this matters: Observability is the missing layer for agentic AI. Without it, autonomy is fragile: small errors silently compound, deployments become risky, and engineering cost goes up because failures aren’t reproducible or attributable.

Who this is for: 1. Teams shipping agentic workflows in production 2. Anyone fighting “unknown why” failures, memory pollution, or runaway context costs 3. Engineers who want auditability + faster debugging loops

Site: https://cortexa.ink/

Would love feedback from anyone running agents at scale: 1.What’s the most painful agent failure mode you’ve seen in production? 2.What signals would you want in an “agent terminal” (retrieval diffs, memory blame, tool-call traces, alerts, etc.)?

Comments

ikharbanda•1h ago
looks like a really cool idea ngl

I Put a Full JVM Inside a Browser Tab

https://bmarti44.substack.com/p/i-put-a-full-jvm-inside-a-browser
1•todsacerdoti•1m ago•0 comments

Full speech pipeline in native Swift/MLX – ASR, TTS, speech-to-speech, on-device

https://github.com/ivan-digital/qwen3-asr-swift
1•ipotapov•1m ago•1 comments

People in northeast BC say rest of province should embrace year-round time zone

https://www.cbc.ca/news/canada/british-columbia/time-change-british-columbia-9.7112139
1•divbzero•2m ago•0 comments

California to require age verification for all OS including Linux

https://www.tomshardware.com/software/operating-systems/california-introduces-age-verification-law
1•hambes•2m ago•0 comments

Rare Not Random – Using Token Efficiency for Secrets Scanning

https://lookingatcomputer.substack.com/p/rare-not-random
1•boyter•4m ago•0 comments

Strict Monospace Font for Chinese Japanese Korean for AI LLM-CLI Users|CodexMono

https://www.npmjs.com/package/@monolex/codexmono
1•monokist•7m ago•1 comments

Working on Things That Suck

https://mayberay.bearblog.dev/working-on-things-that-suck/
1•mugamuga•10m ago•0 comments

Ask HN: How Do Emergency Alerts on Phone Work?

2•rishikeshs•15m ago•1 comments

US President struggles to explain why he launched another Middle Eastern war

https://www.ft.com/content/fd31c6ad-39f0-4fae-851c-fadf44f006eb
6•Jimmc414•21m ago•3 comments

Apple Does Value (Week)

https://om.co/2026/03/02/apple-does-value-week/
1•tosh•22m ago•1 comments

The Pointless War Between The Pentagon and Anthropic

https://www.wsj.com/opinion/the-pointless-war-between-the-pentagon-and-anthropic-9284fd37
3•jrosenblatt•29m ago•0 comments

Show HN: wo; a better CD for repo management

https://github.com/anishalle/wo
1•itsagamer124•32m ago•0 comments

Show HN: AI gaming copilot that uses a phone camera instead of screen capture

https://github.com/ninja-otaku/Project_Aegis
1•Genome123•33m ago•0 comments

OpenAI Amends A.I. Deal with The Pentagon

https://www.nytimes.com/2026/03/02/technology/openai-pentagon-deal-amended-surveillance.html
2•fatboy•33m ago•1 comments

Show HN: Archilvx-Own your Twitter data because cloud tools will fail you

https://www.archivlyx.com/twitter-archive
1•ErinSunny•35m ago•0 comments

Israel hacked Iran traffic cams for years to pinpoint Khaemnei prior to strike

https://www.google.com/search?q=https://www.ft.com/content/1317d740-410c-46a2-97b7-6573e0477121
3•c420•37m ago•1 comments

What 10 Years of Building Social Apps Taught Me

https://twitter.com/nikitabier/status/1481118406749220868
2•metmirr•38m ago•0 comments

OSINT Expert Recreates Timeline of Operation Epic Fury in Interactive 3D Model

https://www.jfeed.com/news-world/operation-epic-fury-3d-replay
2•rapnie•42m ago•1 comments

Show HN: Trade Stocks and Crypto On-Chain with Full Transparency

https://www.aulico.com
2•rendernos•43m ago•0 comments

Buckle Up for Bumpier Skies

https://www.newyorker.com/magazine/2026/03/09/buckle-up-for-bumpier-skies
9•littlexsparkee•51m ago•1 comments

Claude's Constitution and Asimov's Laws

https://yadin.com/notes/asimov/
3•dryadin•1h ago•0 comments

Data-structure-typed – TreeMap, Heap, Graph and more for TypeScript

https://github.com/zrwusa/data-structure-typed
1•zrwusa•1h ago•0 comments

Stop macOS 26 nagging with one tiny policy tweak

https://www.theregister.com/2026/03/02/stop_tahoe_update/
2•fghorow•1h ago•0 comments

Is Mandarin superior for LLM data?

https://medium.com/@tjanmichela/the-language-of-intelligence-could-mandarin-be-the-secret-to-smar...
4•treebeard901•1h ago•3 comments

Show HN: DiffMem in production, Git-based AI memory

https://withanna.io
2•alexmrv•1h ago•0 comments

Constraints and the Lost Art of Optimization

https://denodell.com/blog/constraints-and-the-lost-art-of-optimization
1•CharlesW•1h ago•0 comments

Show HN: Calendar Tool for College Students

https://almanaccal.com/
1•TG_Dev•1h ago•0 comments

Show HN: Private AI Document Server

https://github.com/queryhat/super-hat/blob/main/README.md
1•chhetri978•1h ago•1 comments

Underrated Reasons to Dislike AI

https://www.autodidacts.io/underrated-reasons-to-dislike-ai/
1•Curiositry•1h ago•0 comments

Show HN: Word Doodle – A generative engine that turns text into dense doodle art

https://j-ncel.github.io/word-doodle/
2•koalux•1h ago•1 comments