Hacker News: zopper

New comment by zopper in "Beam: Reflection's 501B open-weight model"

zopper — Mon, 05 Oct 2026 19:59:39 +0000

Open model that is not yet open or widely accessible via API. Primarily comparing to non-SOTA models like Inkling and GLM 5.2. Included comparison to GLM 5.3 and DeepSeek V4.1 Flash in the table, but not in the charts (I assume they would make them look bad). Also no results from AA Index or Arena.

New comment by zopper in "Show HN: StartupWiki – A Free Alternative to Crunchbase"

zopper — Sun, 21 Jun 2026 09:09:53 +0000

Really cool concept but so much of the data is wrong. Anthropic ARR is an order of magnitude higher, Replicate did a Series B as well which is not mentioned. There is probably a lot more.

New Gemini model significantly outperforms others on Chatbot Arena (LMSYS)

zopper — Fri, 06 Dec 2024 17:18:25 +0000

Article URL: https://lmarena.ai/

Comments URL: https://news.ycombinator.com/item?id=42341788

Points: 110

# Comments: 18

New comment by zopper in "The Role of Anchor Tokens in Self-Attention Networks"

zopper — Fri, 11 Oct 2024 17:39:40 +0000

Surprised this isn't getting more attention. It is one of those papers that is very elegant and simple, yet very effective.

New Gemini 1.5 Pro (0801) top on LMSys leaderborad

zopper — Thu, 01 Aug 2024 16:44:24 +0000

Article URL: https://twitter.com/lmsysorg/status/1819048821294547441

Comments URL: https://news.ycombinator.com/item?id=41131021

Points: 9

# Comments: 0

New comment by zopper in "Show HN: I made an app that helps you find where to stream movies and TV shows"

zopper — Wed, 01 May 2024 07:42:44 +0000

Seems like the best way to monetize this. Doubt anyone will regularly pay for the website and banner ads are annoying. Buying a VPN is the logical next step if you find a movie in a different country, might be nice to have multiple affiliate links but this seems like a win-win, it makes it easier to get a VPN for someone who isn't too familiar and the creator earns a bit for their work.

New comment by zopper in "DBRX: A new open LLM"

zopper — Wed, 27 Mar 2024 19:09:02 +0000

Interesting that they haven't release DBRX MoE-A and B. For many use-cases, smaller models are sufficient. Wonder why that is?

GemMoE: An 8x8 Mixture Of Experts based on Gemma

zopper — Wed, 13 Mar 2024 18:24:08 +0000

Article URL: https://huggingface.co/Crystalcareai/GemMoE-Beta-1

Comments URL: https://news.ycombinator.com/item?id=39695319

Points: 4

# Comments: 1

New comment by zopper in "Web Scraping in Python – The Complete Guide"

zopper — Tue, 20 Feb 2024 17:32:04 +0000

This guide (and most other guides) are missing a massive tip: Separate the crawling (finding urls and fetching the HTML content) from the scraping step (extracting structured data out of the HTML).

More than once, I wrote a scraper that did both of these steps together. Only later I realized that I forgot to extract some information that I need and had to do the costly task of re-crawling and scraping everything.

If you do this in two steps, you can always go back, change the scraper and quickly rerun it on historical data instead of re-crawling everything from scratch.

New comment by zopper in "Ask HN: How are you using AI at your company?"

zopper — Fri, 09 Feb 2024 15:24:45 +0000

I'm using a multi-modal LLM to extract floor sizes from property descriptions and floorplans. Working to expand this to other information such as the floor level, if the property has an open plan kitchen etc. (Working on flats.fyi)

New comment by zopper in "Distill OpenAI GPT-4 to a cheaper, faster GPT-3.5 model"

zopper — Fri, 09 Feb 2024 14:38:51 +0000

Looks really interesting but its a shame that it's primarily built around text only. There has been a lot of research around multi-modal LLMs, any plans to support those?

New comment by zopper in "Show HN: Medical LLM API on par with Google Med-PaLM 2. 92% USMLE accuracy"

zopper — Mon, 07 Aug 2023 18:35:07 +0000

This looks really cool. Does your API also support follow up questions and a longer conversations?

New comment by zopper in "Ask HN: If you use an LLM, which one do you use?"

zopper — Tue, 30 May 2023 06:34:29 +0000

Do you run Vicuna locally or have you had any success in deploying it somewhere? I have been thinking of using a fine-tuned model but haven't found anything cost effective where I cab deploy it. You either run the machine where it's served consistently and the cost is in the hundreds or you accept extremely slow starts.

New comment by zopper in "Ask HN: If you use an LLM, which one do you use?"

zopper — Tue, 30 May 2023 06:26:16 +0000

Why did you decide to use both OpenAI and Bard?

Ask HN: If you use an LLM, which one do you use?

zopper — Mon, 29 May 2023 20:49:18 +0000

It seems to me that most startups use one of the LLMs from OpenAI (ChatGPT / GPT-4), other closed-source LLMs like Bard or Claude are not that popular and open-source LLMs are only really used by researchers or hobbyists who run them locally.

Curious to hear from others who are building something on top of an LLM, which one do you use? Has anyone fine-tuned and deployed their own LLM or do you just rely on ChatGPT?

Comments URL: https://news.ycombinator.com/item?id=36117969

Points: 7

# Comments: 7

New comment by zopper in "Show HN: WebscrapeAI – a no-code tool for webscraping"

zopper — Thu, 25 May 2023 07:22:22 +0000

Shame you have to enter your credit card details before even being able to try it. This discouraged me from even signing up.

New comment by zopper in "Ask HN: Has anyone switched from a professional job to a more manual one?"

zopper — Wed, 24 May 2023 13:58:46 +0000

My cousin worked in finance but then due to a health issues he was not able to look at screens (or artificial light in general) for extended periods of time.

He was forced to switch to another profession which had to be outdoor based to avoid artificial light and doesn't use any screens.

He decided to train dogs. Initially it was a bit rough (working long hours a day Monday to Saturday) but fast forward 2 years later he is now considered to be one of the best dog trainers in his area. This allowed him to triple his prices and reduce his working hours. He now earns more than he did in his finance job working shorter hours.

If you find your niche and become an expert in that area, I think you can earn a good living no matter what that niche is. But some niches make it easier than others as the base earnings are higher.

New comment by zopper in "Programmers who learned how to code in school have an edge?"

zopper — Wed, 24 May 2023 13:26:51 +0000

I have found that self-taught programmers are better at debugging as they frequently had no one to turn to when learning as compared to someone who started studying in university. They are frequently also more passionate about programming than someone who just studied at school to get good grades or to secure a job.

But agree with other comments that if you are completely self-taught you will likely miss some fundamental knowledge which is crucial when you go from working on small side-projects to building large systems.

New comment by zopper in "Launch HN: Deepnote (YC S19) – A better data science notebook"

zopper — Fri, 30 Oct 2020 15:53:59 +0000

Loving Deepnote so far. Being able to seamlessly collaborate with others during the pandemic is a lifesaver!

It would be great if you had automatic versioning similar to Google Docs, using Git with notebooks is a nightmare.