frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Google in Your Terminal

https://gogcli.sh/
1•johlo•1m ago•0 comments

Shannon: Claude Code for Pen Testing

https://github.com/KeygraphHQ/shannon
1•hendler•1m ago•0 comments

Anthropic: Latest Claude model finds more than 500 vulnerabilities

https://www.scworld.com/news/anthropic-latest-claude-model-finds-more-than-500-vulnerabilities
1•Bender•6m ago•0 comments

Brooklyn cemetery plans human composting option, stirring interest and debate

https://www.cbsnews.com/newyork/news/brooklyn-green-wood-cemetery-human-composting/
1•geox•6m ago•0 comments

Why the 'Strivers' Are Right

https://greyenlightenment.com/2026/02/03/the-strivers-were-right-all-along/
1•paulpauper•7m ago•0 comments

Brain Dumps as a Literary Form

https://davegriffith.substack.com/p/brain-dumps-as-a-literary-form
1•gmays•7m ago•0 comments

Agentic Coding and the Problem of Oracles

https://epkconsulting.substack.com/p/agentic-coding-and-the-problem-of
1•qingsworkshop•8m ago•0 comments

Malicious packages for dYdX cryptocurrency exchange empties user wallets

https://arstechnica.com/security/2026/02/malicious-packages-for-dydx-cryptocurrency-exchange-empt...
1•Bender•8m ago•0 comments

Show HN: I built a <400ms latency voice agent that runs on a 4gb vram GTX 1650"

https://github.com/pheonix-delta/axiom-voice-agent
1•shubham-coder•9m ago•0 comments

Penisgate erupts at Olympics; scandal exposes risks of bulking your bulge

https://arstechnica.com/health/2026/02/penisgate-erupts-at-olympics-scandal-exposes-risks-of-bulk...
3•Bender•9m ago•0 comments

Arcan Explained: A browser for different webs

https://arcan-fe.com/2026/01/26/arcan-explained-a-browser-for-different-webs/
1•fanf2•11m ago•0 comments

What did we learn from the AI Village in 2025?

https://theaidigest.org/village/blog/what-we-learned-2025
1•mrkO99•11m ago•0 comments

An open replacement for the IBM 3174 Establishment Controller

https://github.com/lowobservable/oec
1•bri3d•14m ago•0 comments

The P in PGP isn't for pain: encrypting emails in the browser

https://ckardaris.github.io/blog/2026/02/07/encrypted-email.html
2•ckardaris•16m ago•0 comments

Show HN: Mirror Parliament where users vote on top of politicians and draft laws

https://github.com/fokdelafons/lustra
1•fokdelafons•16m ago•1 comments

Ask HN: Opus 4.6 ignoring instructions, how to use 4.5 in Claude Code instead?

1•Chance-Device•18m ago•0 comments

We Mourn Our Craft

https://nolanlawson.com/2026/02/07/we-mourn-our-craft/
1•ColinWright•21m ago•0 comments

Jim Fan calls pixels the ultimate motor controller

https://robotsandstartups.substack.com/p/humanoids-platform-urdf-kitchen-nvidias
1•robotlaunch•24m ago•0 comments

Exploring a Modern SMTPE 2110 Broadcast Truck with My Dad

https://www.jeffgeerling.com/blog/2026/exploring-a-modern-smpte-2110-broadcast-truck-with-my-dad/
1•HotGarbage•24m ago•0 comments

AI UX Playground: Real-world examples of AI interaction design

https://www.aiuxplayground.com/
1•javiercr•25m ago•0 comments

The Field Guide to Design Futures

https://designfutures.guide/
1•andyjohnson0•26m ago•0 comments

The Other Leverage in Software and AI

https://tomtunguz.com/the-other-leverage-in-software-and-ai/
1•gmays•27m ago•0 comments

AUR malware scanner written in Rust

https://github.com/Sohimaster/traur
3•sohimaster•30m ago•1 comments

Free FFmpeg API [video]

https://www.youtube.com/watch?v=6RAuSVa4MLI
3•harshalone•30m ago•1 comments

Are AI agents ready for the workplace? A new benchmark raises doubts

https://techcrunch.com/2026/01/22/are-ai-agents-ready-for-the-workplace-a-new-benchmark-raises-do...
2•PaulHoule•35m ago•0 comments

Show HN: AI Watermark and Stego Scanner

https://ulrischa.github.io/AIWatermarkDetector/
1•ulrischa•35m ago•0 comments

Clarity vs. complexity: the invisible work of subtraction

https://www.alexscamp.com/p/clarity-vs-complexity-the-invisible
1•dovhyi•36m ago•0 comments

Solid-State Freezer Needs No Refrigerants

https://spectrum.ieee.org/subzero-elastocaloric-cooling
2•Brajeshwar•37m ago•0 comments

Ask HN: Will LLMs/AI Decrease Human Intelligence and Make Expertise a Commodity?

1•mc-0•38m ago•1 comments

From Zero to Hero: A Brief Introduction to Spring Boot

https://jcob-sikorski.github.io/me/writing/from-zero-to-hello-world-spring-boot
1•jcob_sikorski•38m ago•1 comments
Open in hackernews

Universal pre-training by iterated random computation

https://arxiv.org/abs/2506.20057
37•liamdgray•7mo ago

Comments

liamdgray•7mo ago
Abstract: "We investigate the use of randomly generated data for the sake of pre-training a model. We justify this approach theoretically from the perspective of algorithmic complexity, building on recent research that shows that sequence models can be trained to approximate Solomonoff induction. We derive similar, but complementary theoretical results. We show empirically that synthetically generated data can be used to pre-train a model before the data is seen. We replicate earlier results that models trained this way show zero-shot in-context learning across a variety of datasets, and that this performance improves with scale. We extend earlier results to real-world data, and show that finetuning a model after pre-training offers faster convergence and better generalization."
bionhoward•7mo ago
This is a cool concept, but for comparison, I can’t help but wish there was more comparison between the treatment group and a control group that doesn’t see any universal pretraining data.

It’s good to compare various model sizes and evaluation tasks and random data generators. I just think the paper would more effectively prove its point if it could show models of same sizes which see this random data can learn better from evaluation data later on.

Could even take the initial checkpoint of the model before universal pretraining against the pretrained checkpoint. If the method works, the one that did UP will win.

Maybe I’m way off, I’ll admit I only skimmed it so far. Seems promising, just wishing for some controls.

yorwba•7mo ago
In figures 2, 4, and 6, the top left end of the training curves represents models that have not seen any pretraining data. In figure 5, they're represented by dashed curves.
visarga•7mo ago
Results are modest, maybe 20-30% fewer training steps to reach target performance. This won't solve the problem of organic data exhaustion. We need 100x more data.

They didn't test against actual language model pretraining, only tested against a random init.

- A: Pre-trained on their synthetic LSTM data -> fine-tuned on Wikipedia

- B: Pre-trained on different natural language corpus -> fine-tuned on Wikipedia

- C: Random initialization -> fine-tuned on Wikipedia

They only test A vs C, not A vs B.

WithinReason•7mo ago
This paper addresses the problem of running out of data. You can't do B when you ran out of data so it's irrelevant.
impossiblefork•7mo ago
20-30% isn't modest. I think there is a big problem though, but it's that it's character level prediction.

It's not obvious how generate this kind of good synthetic data when it's to be fed to a tokenized model.