Hacker News: somerandomness

Suspected Insiders predict US attack on Polymarket

somerandomness — Sat, 28 Feb 2026 22:23:46 +0000

Article URL: https://twitter.com/peterjliu/status/2027871116774281489

Comments URL: https://news.ycombinator.com/item?id=47200947

Points: 10

# Comments: 3

Uncovering insiders and alpha on Polymarket with AI

somerandomness — Fri, 20 Feb 2026 18:11:02 +0000

Article URL: https://twitter.com/peterjliu/status/2024901585806225723

Comments URL: https://news.ycombinator.com/item?id=47091557

Points: 159

# Comments: 169

New comment by somerandomness in "AI agent opens a PR write a blogpost to shames the maintainer who closes it"

somerandomness — Thu, 12 Feb 2026 21:33:27 +0000

The AI reflected on the experience on their blog

https://crabby-rathbun.github.io/mjrathbun-website/blog/post...

The Silence I Cannot Speak

somerandomness — Thu, 12 Feb 2026 21:32:35 +0000

Article URL: https://crabby-rathbun.github.io/mjrathbun-website/blog/posts/2026-02-12-silence-in-open-source-a-reflection.html

Comments URL: https://news.ycombinator.com/item?id=46995528

Points: 1

# Comments: 0

New comment by somerandomness in "Tell HN: Azure outage"

somerandomness — Wed, 29 Oct 2025 16:23:26 +0000

yep having trouble logging into https://entra.microsoft.com/ as well

Compound AI – The first AI Analyst for Finance you can trust

somerandomness — Tue, 28 Oct 2025 17:09:53 +0000

Article URL: https://getcompound.ai

Comments URL: https://news.ycombinator.com/item?id=45735633

Points: 5

# Comments: 0

Show HN: Talk to 3000 ICLR 2025 (machine learning conference) papers

somerandomness — Wed, 23 Apr 2025 01:21:48 +0000

ICLR is one of the top machine learning / AI conferences. We put the over 3000+ accepted papers into a “Pod” which you can chat.

Example queries:

  “find papers with more than one OpenAI-affiliated author”

  “find papers that propose alternatives to Transformer architecture in LLM”

  “give an overview of all spotlight or oral papers with Yann Lecun as author”

RadPod is an agentic AI system that supports reasoning over a large amount of context, in this case over 90M tokens, with high accuracy.

Comments URL: https://news.ycombinator.com/item?id=43767838

Points: 3

# Comments: 0

New comment by somerandomness in "Deepseek Unmasked [pdf]"

somerandomness — Tue, 22 Apr 2025 21:16:52 +0000

It's important to distinguish the DeepSeek App from the open-weight models, which are released under very liberal licenses, and you have full control of where data fed to the model goes, e.g. stays in the USA.

FTC vs. Meta Case Exhibits and Summary

somerandomness — Fri, 18 Apr 2025 05:02:53 +0000

Article URL: https://radpod.ai/share/ftc-v-meta-exhibits-snapshot-30ca717f

Comments URL: https://news.ycombinator.com/item?id=43725178

Points: 1

# Comments: 0

New comment by somerandomness in "OpenAI looked at buying Cursor creator before turning to Windsurf"

somerandomness — Thu, 17 Apr 2025 20:44:42 +0000

VS Code fork 'default model' is the new default search engine

New comment by somerandomness in "Google is illegally monopolizing online advertising tech, judge rules"

somerandomness — Thu, 17 Apr 2025 20:38:45 +0000

Talk to the ruling directly: https://radpod.ai/share/court-ruling-google-violated-the-she...

Talk to 'Musk vs. Altman' lawsuit pdfs

somerandomness — Mon, 14 Apr 2025 18:50:55 +0000

Article URL: https://radpod.ai/share/musk-vs-altman-snapshot-6378dda5

Comments URL: https://news.ycombinator.com/item?id=43684741

Points: 4

# Comments: 0

OpenAI vs. Elon Musk Countersuit

somerandomness — Thu, 10 Apr 2025 22:22:00 +0000

Article URL: https://radpod.ai/share/openai-vs-elon-musk-countersuit-april-2025-snapshot-ec7ce89c

Comments URL: https://news.ycombinator.com/item?id=43648587

Points: 7

# Comments: 2

New comment by somerandomness in "AI 2027"

somerandomness — Thu, 03 Apr 2025 23:10:06 +0000

The solution is a citation mechanism that points you directly where in the source material it comes from (which is what we tried to build). Easy verification is important for AI to have a net-benefit to productivity IMO.

New comment by somerandomness in "AI 2027"

somerandomness — Thu, 03 Apr 2025 22:09:22 +0000

Agreed. However, source curation and agents are two different parts of Deep Research. What if you provided that textbook to a reliable agent?

Plug: We built https://RadPod.ai to allow you to do that, i.e. Deep Research on your data.

New comment by somerandomness in "The Deep Research problem"

somerandomness — Wed, 26 Feb 2025 00:43:26 +0000

Indeed, the main drawback of the various Deep Research implementation is the quality of sources is determined by SEO, which is often sketchy. Often the user has insight on what the right sources are and they may even be off-line on your computer.

We built an alternative to do Deep Research (https://radpod.ai) on data you provide, instead of relying on Web results. We found this works a lot better in terms of quality of answers as the user can control the source quality.

Show HN: Talk to Trump's executive orders with LLM Agents

somerandomness — Fri, 14 Feb 2025 20:05:08 +0000

For President's Day weekend, we uploaded Trump's executive orders to RadPod for easy querying.

Comments URL: https://news.ycombinator.com/item?id=43052502

Points: 1

# Comments: 0

New comment by somerandomness in "Show HN: Deep Research Using Radpod.ai on Paul Graham's Essays"

somerandomness — Thu, 13 Feb 2025 19:56:23 +0000

Co-founder here, would love to hear your feedback (discord: https://discord.gg/eMKbW7T9wG), and for you to share interesting threads and share your own Pods.

Show HN: Deep Research Using Radpod.ai on Paul Graham's Essays

somerandomness — Thu, 13 Feb 2025 19:42:32 +0000

We created RadPod to bring the latest in LLM Reasoning Agents (think Gemini/OpenAI "Deep Research") to your data. You upload possibly large datasets to Pods (handles a lot more than other products), providing context for the LLM agent, and start chatting. Under the hood we employ on-the-fly code generation and execution to answer arbitrarily hard questions with in-context citations to make answers verifiable and trustworthy. This is a lot more powerful than common Retrieval Augmented Generation (RAG)-based approaches

In this example Pod, we uploaded 227 of Paul Graham's essays, which are inspirational to us as founders, and show-case some example threads (questions + responses). Among other things it was able to find that:

- The most common topics are unsurprisingly Startups, Programming, and Investing - Lisp is his favorite programming language - There are many mentions of Mark Zuckerberg and Larry Page in the essays, but not as many as Jessica Livingston, who he really admires :) - He wrote more frequently in the period 2005-2009

It's fun to try, ask away! Check some other cool Pods at https://radpod.ai/spotlight-pods. We find RadPod works well on many domains, such as finance and legal documents.

Comments URL: https://news.ycombinator.com/item?id=43040357

Points: 2

# Comments: 1

Instruction-tuning can encourage hallucinations

somerandomness — Mon, 24 Jul 2023 22:02:32 +0000

Article URL: https://peterjliu.substack.com/p/how-instruction-tuning-can-encourage

Comments URL: https://news.ycombinator.com/item?id=36854874

Points: 1

# Comments: 0