PAndreew

My wife has to submit a bunch of translated documents for procurement processes. Noone really reads those but they are a hard requirement. These are usually in DOCX or sometimes PDF with tables, product images etc. So I built her a little applet which (after clicking two buttons) sends the XML or PDF to an LLM in chunks and it spits back the same structure but with translated text. Of course there are edge cases where it doesn’t work perfectly but I’d say it works 95% of the time in a good enough quality to submit them. Saves her roughly 0.5-1 hour per job and much cheaper than other alternatives.

This. Even if you vibecode stuff and outsource most of the decision making to an LLM, the software has to be maintained over an extended period. Not to mention that running the agent also takes money. And if any of the dependencies (including the LLM itself) goes down and you have no idea how the whole system works, you're cooked.

I hope I'm not the only one who wanted to check PyPi to see what package is this... '-.-

Oh and lame-self-comment, but it's at least partly EU stack: Hetzner and Brevo for hosting and transactional mails, Umami for the non-existing analytics.

As others have said it's just a fraction. I'm in a medium size tech-related company and we have 7500+ in one Github org. We have two orgs, so altogether easily 10K+. Of course most of it is stale, obsolete, sandbox, personal tools, etc. I wouldn't be surprised if Github would have 100K+ internal repos or even more.

Sorry for my ignorance, but then couldn't we build this into NPM itself? So before a package is publicly available it would be quaranteened and checked.

I’m running a local Whisper + Gemma 4 pipeline with a cheap USB mic to extract health related data and potential todos from ambient speech. It doesn’t have to be fast doesn’t have to be 100% correct because if it captures at least a few bits of interesting information that would otherwise go unnoticed it’s still a win.

I'm starting to feel like a parrot, but people seem to forget that software engineering is actually a very narrow slice of the white collar pie. You don't need a mega-model which can reason about 100 000 lines of code when you want to create a nice PPT (which consumed literally hours of your life before) to impress your boss. SOTA models will probably be used for frontier research, complex coding tasks, large scale data analysis, etc. And the average Joe shall be able to buy a pre-configured box with a plug-and-play harness and run medium models air-gapped. Or use such models through cloud APIs dirt cheap if privacy is not a concern.

I was thinking about this and there are several aspects that can still make this viable. 1) AI labs are incentivised to increase token consumption because literally that's their product. The only thing they sell AFIAK are tokens (and maybe a teensy bit of user data). So if you build a product that is actively reducing token consumption (which they simply cannot do without hurting themselves even if their marketing fluff says otherwise) you'll save large amounts of money for your customers and they'll choose you. 2) Big providers want to funnel every prompt into their servers. If you're in a regulated market or simply don't want to share every detail with an American or Chinese megacorp you are in trouble. BUT open weight models are now quite capable for "small business stuff" and they can be self hosted. If you can bundle this into your service, in other words actually care about their privacy, they will choose you. Even more so if you're in Europe.

The only place I'd ever talk to a machine is my car. Instead of huge flashy screens that distracts and kills thousands of people maybe they could build a buttons + voice agent system that could actually be useful and durable. I hate to tap Waze/Maps/etc. every time when I go somewhere or that I cannot comfortably switch to specific songs en route without risking my life...