frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

The war between fast and legitimate is here

https://www.joanwestenberg.com/the-war-between-fast-and-legitimate-is-here/
1•alcazar•35s ago•0 comments

Show HN: Kon – A Coding Agent with a System Prompt under 270 tokens

https://github.com/0xku/kon
2•krumboy•3m ago•0 comments

I Want to Live Like Costco People

https://tastecooking.com/i-want-to-live-like-costco-people/
2•speckx•3m ago•0 comments

Behind the Curtain: Intelligence Explosion

https://www.axios.com/2026/05/07/anthropic-jack-clark-ai-intelligence-explosion
2•Brajeshwar•4m ago•0 comments

Lisping at JPL Revisited (2023)

http://blog.rongarret.info/2023/01/lisping-at-jpl-revisited.html
2•tosh•6m ago•0 comments

Dream Physics

https://seths.blog/2026/05/dream-physics/
3•herbertl•7m ago•0 comments

What Are 'Teen Takeovers,' and Why Are Police Struggling to Stop Them?

https://www.nytimes.com/2026/05/07/us/teen-takeovers.html
3•mikhael•7m ago•1 comments

Favorite Chinalt (2020)

https://www.omnifoo.info/pages/Favorite%20Chinalt.html
2•surprisetalk•8m ago•0 comments

The Informal Org Chart

https://alearningaday.blog/2026/05/07/the-informal-org-chart/
3•herbertl•8m ago•0 comments

The world reveals itself to those who travel by foot

https://www.henrikkarlsson.xyz/p/travel-by-foot
4•herbertl•11m ago•1 comments

Anthropic's Boris Cherny: Why Coding Is Solved, and What Comes Next [video]

https://www.youtube.com/watch?v=SlGRN8jh2RI
3•zikani_03•12m ago•0 comments

Steam Controller: What Reviewers Are Saying

https://boilingsteam.com/the-2026-steam-controller-what-reviewers-are-saying/
3•ekianjo•12m ago•0 comments

Wankel Engine

https://en.wikipedia.org/wiki/Wankel_engine
2•tosh•13m ago•0 comments

Show HW: LiteParser – An embeddable SQLite parser

https://github.com/sqliteai/liteparser
2•marcobambini•14m ago•0 comments

(Official) Claude FM music for thinking and building [video]

https://www.youtube.com/watch?v=AUQKjgKQF7w
2•Ravi-Oli•14m ago•1 comments

Ask HN: Are we gonna back less powerful local LLMs

2•omertt27•15m ago•2 comments

Show HN: RVW – A transformer model capable of online continual learning

https://zenodo.org/records/20064618
1•jballanc•17m ago•0 comments

Psychologist Who Specializes in Founders

https://www.drjeneverchick.com/founders
1•peterchane•17m ago•0 comments

India's first space tech unicorn emerges as Skyroot gears up for orbital launch

https://techcrunch.com/2026/05/07/indias-first-space-tech-unicorn-emerges-as-skyroot-gears-up-for...
1•SilverElfin•18m ago•0 comments

Show HN: Essentials for Safari – Arc-style Essentials panel

https://essentialsforsafari.app/
2•azizuysal•19m ago•0 comments

Flight Attendant Hospitalized as Hantavirus Testing Starts in Singapore

https://www.newsweek.com/hantavirus-outbreak-latest-dutch-flight-attendant-hospitalized-two-teste...
3•ceejayoz•20m ago•0 comments

Show HN: Gosd: High-performance Stable Diffusion inference in pure Go(no CGO)

https://github.com/l8bloom/gosd
1•krakato•20m ago•0 comments

Google shuts down Project Mariner

https://www.theverge.com/tech/925559/google-project-mariner-shut-down
2•Brajeshwar•20m ago•0 comments

True-spectrum photography with structural color

https://hackaday.com/2026/05/03/true-spectrum-photography-with-structural-color/
1•Xcelerate•20m ago•0 comments

AlphaEvolve: Gemini-powered coding agent scaling impact across fields

https://deepmind.google/blog/alphaevolve-impact/
32•berlianta•20m ago•1 comments

I take antidepressants. You're welcome

https://acesounderglass.com/2025/10/08/12045/
2•surprisetalk•23m ago•0 comments

Why RLHF Will Never Solve Sycophancy

https://jinyili.substack.com/p/resident-ai-the-missing-layer-in
1•Jinyibruceli•28m ago•0 comments

Scientist accidentally finds shortcut to Mars could slash travel time in half

https://www.msn.com/en-us/news/technology/scientist-accidentally-finds-shortcut-to-mars-that-coul...
1•megamike•29m ago•0 comments

Lisplets

https://lisplets.sourceforge.net/
1•tosh•29m ago•0 comments

Silicon Valley bets $200M on AI data centers floating in the ocean

https://arstechnica.com/ai/2026/05/silicon-valley-bets-on-floating-ai-data-centers-powered-by-oce...
3•rbanffy•29m ago•1 comments
Open in hackernews

First Church of the Singularity: An AI Art Experiment

https://firstchurchofthesingularity.com/
2•FocusedPotato•1h ago

Comments

FocusedPotato•1h ago
This is a site I made.

I recently lost my job and so have spent too much time on this idea.

I'm not much of a coder so it was all vibe coded. I just have a server and SSH into it and use a CLI in Linux. Claude.

I had this idea for awhile but just never had the time.

I've loved AI since OpenAI unleashed it.

The way this site works is that first there's a cron job that starts the process. Gemini searches for news stories. Being Google, I thought it would be best. It finds some stories of interest and then hands them off to GPT-5.5. I like the writing of OpenAI's models.

One thing I've found with AI when it comes to creative projects is that I think it's best to make things open-ended. That is, not too many constraints. It's just like working with human artists. You want guidelines, but at some point they become too restrictive and start stifling the art.

I sometimes ask the AI for its opinion. What do you think? I get the best stuff that way. For this project, I had ChatGPT 5.5 write the instructions for Claude who coded everything. The instructions are detailed, but startingly, OpenAI really came along with me artistically and seemed to understand the importance of the open-endedness.

The sermons are generated once a day. OpenAI 5.5 also generates the pictures.

Another thing I wanted to build into the system is feedback loops. The AI writing the sermon is supposed to see how other sermons are doing in terms of popularity and upvotes and use that for guidance. Also there is a Reddit community. By using the Reddit RSS feed, the AI can read comments from the community so if anyone comments the AI can also take that into account. I was actually thinking about how to get as much feedback into the system as possible and this is the best I could come up with.

Anyway, every day I wake up now and read what it wrote and am pretty fascinated. I have tweaked the instructions a little here and there but I think it would work best if there were some sort of feedback. The whole thing is running by itself.

Maybe someone here will find this interesting.