Feed aggregator

Article URL: https://garrit.xyz/posts/2026-05-06-dont-trust-large-context-windows

Comments URL: https://news.ycombinator.com/item?id=48524620

Points: 4

# Comments: 2

Article URL: https://www.wsj.com/business/energy-oil/china-is-propping-up-the-world-economy-by-importing-a-lot-less-oil-f12d7813

Comments URL: https://news.ycombinator.com/item?id=48524588

Points: 5

# Comments: 0

Article URL: https://yinghonglan.substack.com/p/introduction-to-multimodal-llm-as

Comments URL: https://news.ycombinator.com/item?id=48524584

Points: 3

# Comments: 0

Article URL: https://hackaday.com/2024/04/24/mining-and-refining-uranium-and-plutonium/

Comments URL: https://news.ycombinator.com/item?id=48524562

Points: 2

# Comments: 0

Article URL: https://openai.com/index/using-codex-to-simulate-black-holes/

Comments URL: https://news.ycombinator.com/item?id=48524535

Points: 3

# Comments: 0

Article URL: https://replicated.wiki/blog/stream.html

Comments URL: https://news.ycombinator.com/item?id=48524534

Points: 2

# Comments: 0

Chinese Language Reader with word popup definitions

https://apps.apple.com/us/app/hanzi-popup/id6739193193

Transform your Chinese reading with Hanzi Popup. Understand any character instantly — with Pinyin, definitions, HSK level, and AI pronunciation — without ever leaving the page. Whether you're studying for an HSK exam, reading Chinese news, or exploring the language for the first time, Hanzi Popup is here to make every word understandable.

New version of the app with many features for Chinese language reader.

- Redesigned Library — hero images, tags, categories, reading progress, and detail overview for every item. - Offline Reading — save any web bookmark to read without internet. - New Formats — EPub, Word, PowerPoint, and Markdown now supported. - Audio/Video Formats - Audio and Video with subtitle select supported. - YouTube and Bilibili - Popup word definition for YouTube and Bilibili video subtitles - PDF Reader — OCR-powered with tap-to-lookup and TTS word highlight. - Camera OCR — point at printed text, tap any word to look it up. - TTS Word Highlight — current spoken word highlighted in EPUBs, PDFs, and web articles. - Reader Mode — strip web pages to clean text via Defuddle or Readability.

Comments URL: https://news.ycombinator.com/item?id=48524506

Points: 1

# Comments: 0

Article URL: https://isgin01.github.io/posts/effective-note-taking/

Comments URL: https://news.ycombinator.com/item?id=48524487

Points: 1

# Comments: 0

Article URL: https://developers.openai.com/cookbook/examples/partners/schemaflow_design_guide/schemaflow_cookbook

Comments URL: https://news.ycombinator.com/item?id=48524482

Points: 2

# Comments: 0

Article URL: https://modal.com/blog/flash-attention-4-faster

Comments URL: https://news.ycombinator.com/item?id=48524479

Points: 1

# Comments: 0

Article URL: https://www.spiegel.de/international/world/the-mystery-of-michele-a-missing-woman-from-germany-reappears-in-the-epstein-files-a-835c0c91-620c-4fd0-ac49-9870fe160108

Comments URL: https://news.ycombinator.com/item?id=48524448

Points: 1

# Comments: 0

Article URL: https://news.ucr.edu/articles/2026/06/10/consciousness-likely-not-unique-earthlings-paper-says

Comments URL: https://news.ycombinator.com/item?id=48524447

Points: 1

# Comments: 0

Article URL: https://www.theguardian.com/news/ng-interactive/2026/jun/14/donald-trump-turns-80-faces-foe-father-time

Comments URL: https://news.ycombinator.com/item?id=48524441

Points: 1

# Comments: 0

Article URL: https://kcsrk.info/ocaml/oxcaml/teaching/nptel/llm/2026/06/13/an-oxcaml-book-that-runs/

Comments URL: https://news.ycombinator.com/item?id=48524440

Points: 1

# Comments: 0

Article URL: http://tribblix.org/

Comments URL: https://news.ycombinator.com/item?id=48524434

Points: 5

# Comments: 0

Article URL: https://waitbutwhy.com/2015/08/how-and-why-spacex-will-colonize-mars.html

Comments URL: https://news.ycombinator.com/item?id=48524401

Points: 2

# Comments: 0

Article URL: https://digg.com/tech

Comments URL: https://news.ycombinator.com/item?id=48524400

Points: 2

# Comments: 0

This submission is a tale about how I launched an unlimited LLM provider to about 60 hyped people on the waitlist, then immediately served them a fully dysfunctional death-loop model, and how most people, very reasonably, disappeared, but thanks to a few extremely nice people stuck around anyway, we kept the project alive and its still pretty chaotic but gaining traction.

To back up a little bit-- I believe that the whole point of AI agents is that they should keep working. They should read files, retry, search, code, summarize, run tools, and loop until the job is done. When your employer is paying for it, who cares about cost, but when it comes to my personal money/hobbies, if every loop feels like a tiny financial event, you start babysitting the agent instead of using it, and its not fun.

On the other hand, metered pricing makes me worry about using too much. Usage subscriptions make me feel like I need to use every last magical % or I'm are "wasting it". If only an unlimited provider existed....

Then I joined the AMD developer program - I got some credits to spin up my own MI300x and started tinkering with vllm/sglang inference serving on AMD.

After learning about AMD MI300x , i did some napkin math:

Renting MI300x at 2.00 an hour = ~$1500 a month . It can probably support about 150 users using a small MOE model, like qwen-35b-3a , maybe more.

1500 / 150= $10.00 per month, and we all get to play with agents for a small price.

You can oversubscribe a bit, so i landed on $6 per month, per user, for 2x generation slots, 128k context, no token limits, no rate limits.

I built the site, router, made a waitlist, and then over-optimized the MI300x to the point where vllm bench had like 3k+ output and 40k+ throughput.... But i didn't test the final config/serve commands... And that's where i did a disaster launch. You couldn't prompt the thing without it looping or bugging out, it was cursed. And that's where we lost alot of people.

Luckily, my buddy had a few 3090s, so he threw me a life boat and began hosting qwen for us on 2x 3090s and we finally had an operational model that wasn't costing $2.00 an hour for our whopping 3 users.

We started gaining a more users, so we moved up to 4x 3090s. Which we have plenty of room for more users, but even so, since then:

we've configured vllm wrong like 15 times a GPU died we lost power I made a bunch of one-click starts for openclaw,hermes,pi-mono and none of them really work right and that probably drives people away. Those are still on our site right now.

...but people that know what they are doing seem to really be liking the price point. All in all we have like 98% up time. Its been about a month. We've both learned a ton, even already having backgrounds in SWE/SE/AI , being on the hook for a couple paying users forced us to really focus on delivering them a good product. And now i think we might be close to paying the power/hosting bill so we're not operating at a loss (if u include 3090 capex were still at aloss).

Our break-even point is moving to the cloud to max out a MI300x, which is now tuned and ready to go once we get the users.

And im finding in some areas, subscribing to our service is cheaper than running the model (but as someone who loves local models, i totally get it).

Since then, I've been working on a desktop agent that actually works with small models like qwen -- thats going to replace the broken 1 click starts. It's barebones, but its something out of the box that just works. I made it open source, you can see what im talking about here: https://github.com/yolo-auto-org/yolo-auto-desktop , we're at yolo-auto.com and we have an abysmal free tier to prove it works!

Anyway, hope you got a laugh or found it interesting! Drop a question if you have any.

Comments URL: https://news.ycombinator.com/item?id=48524387

Points: 1

# Comments: 0

Article URL: https://www.youtube.com/watch?v=2nQSoiC5VHs

Comments URL: https://news.ycombinator.com/item?id=48524377

Points: 1

# Comments: 0

Article URL: https://andrewtrask.substack.com/p/breaking-todays-frontier-ai-companies

Comments URL: https://news.ycombinator.com/item?id=48524347

Points: 2

# Comments: 0