Hacker News: geniium

New comment by geniium in "Qwen 3.8 27B"

geniium — Fri, 14 Aug 2026 21:00:33 +0000

I think the model got trained only on pelican to get there

New comment by geniium in "DeepSeek V4 Flash 0731"

geniium — Sat, 08 Aug 2026 05:46:17 +0000

It's really amazing to see how the gaps between the self hostable models and the closed models has been shrinking in the last 24 months.

And how this has been accelerating!!

I felt this very hard when I had to travel in the middle of nowhere in south america, with no network, and wanted to keep an LLM model on my macbook pro with 48GB of RAM. That was back in April 2026, a few months ago.

I downloaded Google Gemma 4 (google/gemma-4-26b-a4b) and - Oh boy - I was amazed by it's capacity!

I was able to use it to code simple things, ask it about nature, learn new stuff while traveling and make stories for the kids.

Was really amazing to observe and experiment this!

Seems to me there will be some good chance to run these great LLM locally on our hardware!

Amazing time to be alive

New comment by geniium in "Kimi Work"

geniium — Tue, 21 Jul 2026 15:25:26 +0000

Thanks for making me smile

I love that quote and like to share it often with people around me

New comment by geniium in "Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge"

geniium — Tue, 21 Jul 2026 15:23:58 +0000

Yet it pushes the goal of the brand/shop forward: make the product more appealing and sell

New comment by geniium in "Kimi K3: Open Frontier Intelligence"

geniium — Fri, 17 Jul 2026 10:40:24 +0000

> Chinese labs only have 5-10% the valuation of OpenAI/Anthropic, so massive monopoly profits aren't necessary. Profit expectations for tech companies in China are really low in general, complete opposite of the US.

It’s really amazing to see that the competition is creating better quality models for everyone - and am really happy that some of these are open source (or partially os).

Regarding the valuation, that maybe points a finger to the over valuation of the US companies ?

Interesting times to be alive.

New comment by geniium in "Kimi K3: Open Frontier Intelligence"

geniium — Fri, 17 Jul 2026 10:37:24 +0000

Impressive Pelican, I like it

New comment by geniium in "Show HN: Pulpie – Models for Cleaning the Web"

geniium — Mon, 06 Jul 2026 21:40:24 +0000

Amazing I was just looking for something like this to be able to import web page content into Whisperit

New comment by geniium in "A History of Menus Is a Menu of History"

geniium — Sat, 27 Jun 2026 21:57:26 +0000

So amazing to travel in time with style, congrats to the author

New comment by geniium in "Show HN: Trace – Offline Mac meeting transcripts you can flag mid-call"

geniium — Mon, 15 Jun 2026 01:53:01 +0000

And if it runs on the browser without install it would not probably be able to record your other browser (or app) audio

New comment by geniium in "Hacker News, Sans AI"

geniium — Fri, 05 Jun 2026 21:51:10 +0000

Was about to write the same

New comment by geniium in "Show HN: Smol machines – subsecond coldstart, portable virtual machines"

geniium — Sat, 18 Apr 2026 09:33:32 +0000

Congrats that looks really amazing!

New comment by geniium in "70k Books Found in Hidden Library in This Germany Home (2023)"

geniium — Fri, 06 Mar 2026 20:28:57 +0000

What is very surprising for me is the height of the books. If you look closer at the picture, seems that most of the books are the same height, so that's very strange for any book collection except if he was collecting the same book editors or... Am missing something ?

New comment by geniium in "Discord/Twitch/Snapchat age verification bypass"

geniium — Thu, 12 Feb 2026 03:44:31 +0000

This is the right question. Who will benefits from blocking young people? Probably not the platform.

New comment by geniium in "Luce: First Electric Ferrari"

geniium — Mon, 09 Feb 2026 20:51:33 +0000

yeah, seems EV is a hard market to enter. Porsche seems that have had a hard time entering it too (see numbers, am no expert)

New comment by geniium in "Luce: First Electric Ferrari"

geniium — Mon, 09 Feb 2026 20:50:54 +0000

haha me neither

New comment by geniium in "Testing Ads in ChatGPT"

geniium — Mon, 09 Feb 2026 20:50:32 +0000

This is interesting. On one side, I hate ads everywhere. On the other side, I was always very appreciative of the work of some people like Benetton or some creative ads. I remember watching here in Europe, there is a show that was called Culture Pub every Sunday night that was showing the best ads, and I always enjoyed watching it. That part is probably the best part. The creativity

New comment by geniium in "Testing Ads in ChatGPT"

geniium — Mon, 09 Feb 2026 20:05:59 +0000

Just wanted to mention the ads from Claude here :)

New comment by geniium in "Luce: First Electric Ferrari"

geniium — Mon, 09 Feb 2026 20:04:51 +0000

That doesn't look good. I'm very surprised that a brand like them release such a cheap-ass version.

New comment by geniium in "Ask HN: What are you working on? (February 2026)"

geniium — Mon, 09 Feb 2026 15:37:28 +0000

Of course! I have looked into many tools to be able to do that properly. This one was done with Screen Studio, which is one, and there is another one that is nice and open source: Cap - but a bit less features.

New comment by geniium in "Ask HN: What are you working on? (February 2026)"

geniium — Mon, 09 Feb 2026 12:43:37 +0000

Working on Whisperit v3: a document automation platform for lawyers, voice-first.

The idea: lawyers spend 60%+ of their time writing documents. We give them an AI assistant that actually understands legal context — case files, templates, emails, the full picture.

What we've been shipping lately on the v3 branch:

• Workflow engine — visual editor + execution pipeline for legal document workflows (think: intake → draft → review → send, all orchestrated)

• Case-context AI assistant — docked right into your case view, with file attachments, skill system, and next-action suggestions. It knows your case, not just your prompt. Think Pipedrive CRM for lawyer!

• Skills system — modular prompts lawyers can pick and inject (summarize deposition, draft motion, extract key dates, Claude Skills, etc.)

• Cloud drive sync — bi-directional sync with Google Drive and OneDrive. Lawyers keep their existing file setup, we plug in and keep everything in sync

• Outlook, Gmail, IMAP email integration — connect your inbox, pull relevant emails directly into cases

• Template multi-output generation — one case, multiple documents generated in one flow

• BYO AI provider — user/workspace/admin level settings for AI models. Some firms want Azure, some want Anthropic, we let them choose

• Canvas boards with AI context for visual case planning Stack: Next.js 16, FastAPI + LangGraph backend, Supabase.

We're a small team in Switzerland and across the globe, shipping daily. Target is Q1 for the v3 launch. If you work in legal or know lawyers drowning in document work: would love feedback or test our beta :)

https://whisperit.ai