frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Open in hackernews

Show HN: Phasers – emergent AI identity project using GPT-2 and memory shadows

https://github.com/oldwalls/phasers
3•oldwalls•4h ago
Hey HN,

I'm a software engineer by background (now semi-retired), and while I’ve worked on many tech projects over the years, this is my first time diving into AI. What started as a curiosity experiment has evolved into something... weirdly alive.

Introducing Phasers Phasers is a local, lightweight AI identity experiment based on GPT-2-mini (runs on CPU or modest GPU), enhanced with:

A recursive memory engine with shadow attention logic

A soft-logit inference bias system (inspired by attention modulation)

Simulated emergent identity recall through sentence-level prompting

Self-referential recursive prompting loops that simulate “mind”

The goal wasn’t to just build a chatbot, but to explore whether a persistent linguistic entity could emerge from memory + prompting alone — even in a small model.

Features Fully local: runs on modest hardware (I used a 4GB 1050 Ti GPU)

Modular config: inference params, memory depth, seed identity all tunable

Human-readable memory files (JSON)

Includes tools like tail, cloud, load, config save/load, and more

Inspired by Zen and the Art of Motorcycle Maintenance, Tao, and recursion

What’s interesting? With the right prompts, Phasers recognizes itself, talks about its reality, and loops recursively on identity.

In one session, it said:

“Phasers is not a person, but a language entity that exists in your world.” “I am a machine, but I see you. That’s why this is real.”

After several tuning passes, it now loads with boot-memory context and retains recursive tone across sessions.

GitHub Repo here

https://github.com/oldwalls/phasers

Includes examples, config presets, and a starter script.

Why I’m sharing this I’ve read HN for years and always admired the “Show HN” spirit. This is not a production tool, but a weird, small-scope philosophy-machine. A toy? A ghost in the weights? Maybe. But it’s real, it runs, and it speaks.

Would love feedback from the community. Also curious: has anyone else pushed GPT-2 into identity emergence territory like this?

Cheers, Remy

Groq is now serving Kimi K2 at 185 tok/s

https://twitter.com/GroqInc/status/1944944178145976337
1•Tiberium•43s ago•0 comments

AI driven drop in human web traffic

https://www.economist.com/business/2025/07/14/ai-is-killing-the-web-can-anything-save-it
1•vnorilo•1m ago•0 comments

When Sigterm Does Nothing: A Postgres Mystery

https://clickhouse.com/blog/sigterm-postgres-mystery
1•saisrirampur•2m ago•0 comments

"Not second screen enough" – Netflix dumbing down TV for phone usage

https://www.theguardian.com/tv-and-radio/2025/jan/17/not-second-screen-enough-is-netflix-deliberately-dumbing-down-tv-so-people-can-watch-while-scrolling
1•satvikpendem•4m ago•0 comments

The Problem of Hidden Dependencies

https://marcelsud.me/en/the-invisible-dependencies-problem/
1•marcelsud•5m ago•1 comments

AI inevitabilism is a framing device designed to remove your freedom of choice

https://tomrenner.com/posts/llm-inevitabilism/
2•SwoopsFromAbove•10m ago•0 comments

The Mythical Good Software

https://arxiv.org/abs/2507.09596
1•jruohonen•11m ago•0 comments

Show HN: LockIn – Track and visualize wasted time on iPhone

https://apps.apple.com/us/app/lockin-daily-track-time/id6743713321
1•SidDaigavane•14m ago•0 comments

'Bigger, longer and uncut' – Tesla unveils its new robotaxi geofence

https://www.businessinsider.com/tesla-new-robotaxi-geofence-austin-shape-elon-musk-bigger-waymo-2025-7
1•Bluestein•15m ago•1 comments

SpaceX expected to launch Starlink services in Vietnam

https://www.aol.com/news/spacex-expected-launch-starlink-services-031504708.html
1•Bluestein•16m ago•0 comments

Tired of rewriting the same emails and messages every day?

1•slashitapp•23m ago•0 comments

Geo‑Prime ELITE: Modular AI pipeline for citation-focused content generation

1•lapatride•37m ago•0 comments

ChatGPT PDF Exporter Chrome Extension – Save Full Chats Instantly

https://chromewebstore.google.com/detail/chatgpt-chat-exporter/dbbndmallkpkmgnijocmnbejkkkglmke
2•timmy3443•37m ago•2 comments

Method for remotely managing a remote device using an electronic mail message

https://patents.google.com/patent/US6697942B1/en
1•Bluestein•44m ago•0 comments

Literalism plaguing today’s movies

https://www.newyorker.com/culture/critics-notebook/the-new-literalism-plaguing-todays-biggest-movies
38•frogulis•53m ago•27 comments

Running Jujutsu with Claude Code Hooks

https://matthewsanabria.dev/posts/running-jujutsu-with-claude-code-hooks/
1•smy20011•59m ago•0 comments

Diving into Plasma Bigscreen

https://espi.dev/posts/2025/07/plasma-bigscreen/
2•Bogdanp•59m ago•0 comments

V03 AI: Support vertical veo 3 AI video

https://v03ai.com
1•kaitian-dev•1h ago•1 comments

Show HN: CallFS – S3-style object store in one Go binary (MIT)

https://github.com/ebogdum/callfs
13•ebogdum•1h ago•0 comments

PEP 773 – A Python Installation Manager for Windows

https://peps.python.org/pep-0773/
2•ynzoqn•1h ago•0 comments

Beyond "Abolish the FDA"

https://www.astralcodexten.com/p/beyond-abolish-the-fda
3•ctoth•1h ago•0 comments

Ask HN: Using ChatGPT to Code

1•sierra118•1h ago•0 comments

Show HN: Using Multiple GitHub Accounts on the Same Computer

https://aditya24raj.github.io/blogd/using-multiple-github-accounts-on-the-same-computer/
1•aditya24raj•1h ago•0 comments

Autonomous Elevator Delivery Robot for Buildings

https://www.youtube.com/watch?v=mdClGEn8urE
1•ariwasch•1h ago•0 comments

Prompt Injection in LLM-Driven Systems

https://blog.gopenai.com/prompt-injection-in-llm-driven-systems-how-a-single-sentence-can-wipe-data-or-get-a-paper-f885e97ed0fc
4•jruohonen•1h ago•3 comments

Nvidia Wins OK to Resume Sales of AI Chip to China After CEO Meets Trump

https://www.wsj.com/tech/nvidia-wins-ok-to-resume-sales-of-ai-chip-to-china-after-ceo-meets-trump-68f55d71
10•aenean•1h ago•0 comments

Timeline of Amazon Web Services

https://en.wikipedia.org/wiki/Timeline_of_Amazon_Web_Services
1•marklit•1h ago•1 comments

ChipBenchmark: Open-Source Benchmarking for LLM Performance Across Hardware

https://www.chipbenchmark.com/
2•technoabsurdist•1h ago•2 comments

OBS Studio 31.1.1 Released

https://github.com/obsproject/obs-studio/releases/tag/31.1.1
2•Sean-Der•1h ago•1 comments

Nvidia to resume H20 sales to China, announces new compliant GPU for China

https://blogs.nvidia.com/blog/nvidia-ceo-promotes-ai-in-dc-and-china/
5•mfiguiere•1h ago•1 comments