frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

GPT Guesses Between 1 and 100

https://github.com/exmergo/research-chatgpt-guesses-between-1-and-100
31•adunk•54m ago

Comments

alentodorov•24m ago
ha. and i thought 37signals was pretty random
fny•23m ago
I wonder if Benford's law kicks in with larger numbers.

https://en.wikipedia.org/wiki/Benford%27s_law

penr0se•21m ago
Breaking: language model whose purpose is to predict the most likely token, after being trained on non-uniform human-generated dataset, does not follow a uniform distribution.
singpolyma3•18m ago
What's interesting is not that it isn't random. But rather the particular way in which it isn't random.
vidarh•9m ago
People are also not remotely random in this respect.

See e.g. the "blue 7" phenonmenon [1]. While it is disputed by some, I'ver personally witnessed it "second hand". E.g. before learning of it (I was aware of the general principles of cold reading relying on stats and knowledge of human nature, but not how to do this particular one), a former boss of mine came back from lunch all excited and recounted a guy who'd run a cold reading routine on him that involved the guy getting him to think about blue and 7. Before he got to the answer, I already knew the answer was going to be blue and 7.

[1] https://en.wikipedia.org/wiki/Blue%E2%80%93seven_phenomenon

gruez•20m ago
The topic is vaguely interesting but I stopped reading a few paragraphs in because it's obviously AI generated.
maxloh•18m ago
It could be an attack surface. Maybe one day, when we find a chatbot online, we could let it guess a random number repeatedly, then accurately infer the underlying model based on the resulting distribution.
vidarh•8m ago
At least some Claude models have a thing for numbers that contains "47"...
madanparas•16m ago
bro 42 at 4x. the model read the whole internet and became a Douglas Adams fan.
adrian_b•10m ago
While the results were not surprising, I found interesting that the number "69" was repressed in the output, so not even this kind of mathematical question escapes GPT censorship.

It appears that recognizing the effects of censorship is the easiest way to distinguish answers generated by an "AI' from those generated by a human.

hackinthebochs•5m ago
Also see: https://people.csail.mit.edu/renda/llm-sampling-paper
a3w•4m ago
"69 is a meme number", well no, 69 is innuendo. And sex = bad for bots. 67 is the meme number.

Mnemosyne – Memory for AI Hermes Agents, Sub-Millisecond Recalls, Local First

https://mnemosyne.site/
1•AbdiiSan•1m ago•0 comments

Show HN: Proj – organize your coding projects with categories and one-key CD

1•whizhuii•2m ago•0 comments

Agentic AI Design Patterns for Developers (2026)

https://learnagenticpatterns.com
1•ankitg12•3m ago•0 comments

Local-First Twitter Workspace

https://birdclaw.sh/
1•cat-whisperer•4m ago•0 comments

Beyond Senior: Consider the staff path

https://hawksley.org/2026/01/14/beyond-senior.html
2•RyeCombinator•4m ago•0 comments

Your Function's Doppelgänger (Fenchel Conjugate)

https://fedemagnani.github.io/math/2025/07/04/fenchel.html
1•drunello•5m ago•0 comments

Spec-Drive Development (SDD) compressed with math-glyphs

https://lab5.ca/blog/spec-driven-development/
1•kborovik•6m ago•0 comments

Awesome: Lists about all kinds of interesting topics

https://github.com/sindresorhus/awesome
1•danborn26•8m ago•0 comments

Show HN: I made a compiler/VM for untrusted scripts

https://autolang.vercel.app/docs/philosophy-vision
1•hoansdz•9m ago•0 comments

Nendo's Wonderful Toru, an Electric Kettle for Alessi

https://www.core77.com/posts/143823/Nendos-Wonderful-Toru-an-Electric-Kettle-for-Alessi
1•surprisetalk•10m ago•0 comments

Search engines alternatives now that Google isn't Google anymore

https://techcrunch.com/2026/05/21/six-search-engines-worth-trying-now-that-google-isnt-really-goo...
8•elorant•13m ago•0 comments

Happy Towel Day HN

https://en.wikipedia.org/wiki/Towel_Day
1•salutis•19m ago•0 comments

PEEK: Give Your Agent an Orientation Cache (MIT CSAIL, Khattab group)

https://zhuohangu.github.io/blog-post-peek/
2•galsapir•19m ago•0 comments

Ask HN: Local model experiences with 'high-reasoning distill' finetunes

1•sleepyeldrazi•20m ago•0 comments

Agents Just Need APIs

https://agent-data.dev/blog/benchmarking-ai-agent-web-access/
2•jb_hn•20m ago•0 comments

Velocity in Every Voxel – Perception in Robotics

https://atomsfrontier.substack.com/p/velocity-in-every-voxel
1•jpatel3•21m ago•0 comments

Facebook's Flow is being ported to Rust

https://github.com/facebook/flow/blob/main/rust_port/rust_port_status.md
1•mirekrusin•22m ago•0 comments

Pope Leo warns AI revolution driven by 'idolatry of profit'

https://www.ft.com/content/12313f08-991d-4079-9631-9ce7ae70c3e3
2•1vuio0pswjnm7•23m ago•0 comments

Ask HN: Did agentic coding change the way you think about commit granularity?

1•luodaint•23m ago•0 comments

Show HN: Stumpy – StumbleUpon Re-Created

https://chromewebstore.google.com/detail/stumpy/blfpeiakahiemhdiaaaacfgmdmdocfem
1•postatic•23m ago•0 comments

China launches 'human artificial embryos' into space: off-world reproduction?

https://www.livescience.com/space/space-exploration/china-launches-human-artificial-embryos-to-sp...
1•bookofjoe•23m ago•0 comments

Code Golf

https://code.golf/
1•alabhyajindal•23m ago•0 comments

Unitree Develops Production-Ready Mech Suit

https://www.core77.com/posts/144165/Unitree-Develops-Production-Ready-Mech-Suit
1•surprisetalk•25m ago•0 comments

FuguIta – An OpenBSD-Based Live System

https://fuguita.org/
2•ska80•25m ago•0 comments

The generation vs. verification delta explains why LLM's are useful

https://simianwords.bearblog.dev/the-generation-vs-verification-delta-explains-why-llms-are-useful/
1•simianwords•27m ago•0 comments

The Evolution of Program State – Queue

https://spawn-queue.acm.org/doi/10.1145/3799737
1•rbanffy•27m ago•0 comments

I saw the future of Android Auto, and now Google has me dreading my own car

https://www.techbuzz.ai/articles/i-saw-the-future-of-android-auto-and-now-google-has-me-dreading-...
2•teleforce•31m ago•0 comments

How to Migrate from Vimeo to Bunny Stream

https://bunny.net/blog/how-to-migrate-from-vimeo-to-bunny-stream/
1•m_nalikowski•33m ago•0 comments

The Social Contract of Writing

https://jola.dev/posts/the-social-contract-of-writing
1•shintoist•34m ago•0 comments

Notebooklm-guide.com – 1k Battle-Tested Prompts and 59 Power Workflows (2026)

https://notebooklm-guide.com/
1•ZeroshotCraft•34m ago•0 comments