frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Vexp – Your AI coding agent forgets everything. Mine doesn't

https://vexp.dev
1•nicola_alessi•2h ago
I built vexp because AI coding agents have two expensive problems: they waste tokens reading irrelevant code, and they forget everything between sessions.

The token problem: agents read entire files linearly to build context. On a medium TypeScript project, a single query was consuming ~18k tokens — most of it irrelevant. vexp builds a dependency graph from the AST (who calls what, who imports what, what types flow where) and serves only the relevant subgraph as a token-budgeted capsule. ~2.4k tokens instead of ~18k, with better response quality because the context is precise.

The memory problem: this is where it gets interesting. The obvious approach is giving agents a "save what you learned" tool. They won't use it. I tried every prompting trick. Agents optimize for task completion, not knowledge retention. The incentive structure is fundamentally wrong.

So vexp observes passively. It watches what happens — which symbols the agent explored, which files changed and how they changed structurally, what patterns emerge across sessions — and builds memory without the agent lifting a finger. When code changes, linked memories auto-stale. The agent sees "previous context exists but the code has changed since re-evaluate." It also catches anti-patterns like dead-end exploration and file thrashing so the agent doesn't repeat mistakes.

The memory is hybrid-searched with 5 signals (text relevance, semantic similarity, recency, code graph proximity, staleness) and every result includes a "why" field explaining the ranking. No black box.

Architecture: single native Rust binary (~15MB), SQLite with WAL mode, tree-sitter for 11 languages, MCP protocol. 100% local, zero cloud, zero account, zero network calls. Works with Claude Code, Cursor, Copilot, Windsurf, Zed, Continue, and 6 other agents. Auto-detects which agent is running and generates tailored instruction files.

Free tier: 2k nodes + all memory tools. Everything runs on your machine.

https://vexp.dev

Towards a Science of AI Agent Reliability

https://arxiv.org/abs/2602.16666
1•smartmic•8s ago•0 comments

How we made Docker builds 193x faster across AI agent sessions

https://blog.helix.ml/p/how-we-made-docker-builds-193x-faster
1•quesobob•2m ago•0 comments

Ask HN: Did your client ever replaced you by a more junior freelancer?

1•goingbananas•3m ago•0 comments

Addressing your questions about the Cyber Resilience Act

https://fsfe.org/news/2026/news-20260224-01.html
1•Tomte•3m ago•0 comments

I don't care what tools you use. But – and this is a big but

https://come-from.mad-scientist.club/@algernon/statuses/01KHYGWT17C1HNKRCVBMYTZVHQ
1•latexr•4m ago•0 comments

Show HN: StarkZap – Gasless Bitcoin Payments SDK for TypeScript

https://github.com/keep-starknet-strange/starkzap
1•starkience•4m ago•2 comments

Mercury 2: Diffusion Reasoning Model

https://www.inceptionlabs.ai/blog/introducing-mercury-2
2•zof3•4m ago•0 comments

SpacetimeDB 2.0 [video]

https://www.youtube.com/watch?v=C7gJ_UxVnSk
6•aleasoni•5m ago•0 comments

Show HN: Awsim – Lightweight AWS emulator in Go (40 services in progress)

https://github.com/sivchari/awsim
2•sivchari•6m ago•0 comments

Stripe valued at $159B, 2025 annual letter

https://stripe.com/newsroom/news/stripe-2025-update
2•jez•6m ago•0 comments

The Schema Language Question: The Quest for a Single Source of Truth

https://www.chiply.dev/post-schema-languages
1•chiply•6m ago•0 comments

OpenClaw led to a user's Gmail account being disabled

https://twitter.com/iamlukethedev/status/2025782621066899873
2•idoxer•7m ago•0 comments

Lockrion – a deterministic issuance protocol on Solana (pre-launch)

https://lockrion.com
1•flexionU•8m ago•1 comments

Dangerously Skip Permissions

https://asimovaddendum.substack.com/p/dangerously-skip-permissions
1•srulyrosenblat•9m ago•0 comments

I built a job board that crawls 900 company career pages directly

https://www.jobscroller.net
3•couentine•9m ago•1 comments

Global AI data center boom hits delays

https://www.axios.com/2026/02/24/ai-data-center-boom-projects-numbers
1•1vuio0pswjnm7•9m ago•0 comments

'Probably' doesn't mean the same thing to your AI as it does to you

https://theconversation.com/probably-doesnt-mean-the-same-thing-to-your-ai-as-it-does-to-you-275626
1•geox•9m ago•0 comments

Build your own custom AI CLI with Pydantic AI

https://vstorm-co.github.io/pydantic-deepagents/
2•frankwiles•9m ago•1 comments

AI Homer Simpson 'Cover Songs' Have 'Poisoned' Soulseek

https://www.vice.com/en/article/infinite-ai-homer-simpson-cover-songs-poisoned-soulseek/
2•ides_dev•9m ago•0 comments

QED: John's Not Mad – Documentary 1988 [video]

https://www.youtube.com/watch?v=wxfJDpd3XcY
1•mellosouls•11m ago•1 comments

Single Store Vector Search Index: Architecture and Memory Efficiency

https://memgraph.com/blog/single-store-vector-index
2•mbuda•11m ago•1 comments

Nothing's Happening

https://uptointerpretation.com/posts/nothings-happening/
1•hardwaregeek•12m ago•0 comments

Great Isaiah Scroll, oldest Biblical book ever found, on show for first time

https://www.timesofisrael.com/great-isaiah-scroll-oldest-near-complete-biblical-book-ever-found-o...
1•myth_drannon•12m ago•0 comments

Ask HN: Developers cloned my student project in 2 weeks. Why?

1•glinkswww•12m ago•2 comments

Row Locks with Joins Can Produce Surprising Results in PostgreSQL

https://hakibenita.com/postgres-row-lock-with-join
1•haki•13m ago•0 comments

Show HN: GDBS – Browser geometric physics engine replacing HPC cluster workloads

https://gdbs.getvaultsync.com
1•garrjo•13m ago•0 comments

ISS Space Timelapse сaptures strike flashes over Kyiv

https://twitter.com/RikyUnreal/status/2025965975510655237
1•defly•13m ago•0 comments

I left my AI system running by accident. It was still going 38 hours later

https://github.com/adriancarriger/lambda/blob/main/docs/blog/38-hours.md
1•adriancarriger•14m ago•0 comments

Decision Theory Cuts Your AI Agent's API Bill in Half

https://gfrm.in/posts/decision-theory-agents/index.html
1•slygent•14m ago•0 comments

Show HN: LazyHippo.dev – sandbox for auth and product testing

https://www.lazyhippo.dev/
1•schmommy•14m ago•0 comments