1) Codex Plus (20 bucks) - find it enough even though I'm working on 3 small-medium size projects and occasionally helps out my wife
2) No
3) Pi
4) No, I think these are token burners and makes it practically impossible to track what's happening from a human side.
5) My own tool: <https://github.com/PAndreew/pi_group> - the main benefit is that it's multiplayer so it functions as a regular chat app in which you can invoke Pi in a group chat. This way the agent can have a shared context, but it still have all the features like generative ui, internal browser etc. that other bolted-on solutions like Claude bots in Slack do not possess.
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.
I’m doing something similar at work. Thousands of contracts, need to extract billing terms/logic. It’s scattered throughout each contract but has no summary, each contract can be hundreds of pages of non applicable legalese. The problem is, once I run it, can I trust the output without manual verification? How will I know if it skipped/missed something from a contract?
I would build a review dashboard that shows you the extracted data on one side and the visual snippet of the contract on the other. Quick to double check, and you still save time over opening the contract and searching for the right section and extracting it by hand or via OCR.
We have done something similar where it highlights the terms, basically each extracted term is linked to a specific highlight in the contract, and has them side by side like you said and you can jump around the contract by clicking on the extracted term. This helps when the agent needs HITL clarification on a term, as it gets you there quick and you can tell it more information to aid interpretation. But, if it missing something completely, I’d never know without reading the whole contract and validating every term I found the agent noticed. We’ve done some training sets and it does help and get better but the large variation of contracts means there’s little standardization.
I think we’re on path to just do random full review audits of maybe 10% to get a baseline confidence level and then just note it as a risk. Still a large manual task, we’ll probably outsource it.
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 already submitted my take on this to Show HN: https://currantfeed.cc/. Noone gave a shit so it's most likely a very bad idea and certainly a very steep uphill battle. It's basically an Airbnb for websites that by default randomly sorts them and then you can filter by different attributes. Owners have to submit and maintain their "listings". It does have an optional subscription (I'm not sure if it works haven't tested it).
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.
No OP but I used to work at a large company with a similar number of repos.
When I left about a year ago, we had just started (after being on Github for almost 8 years) an ongoing project of first archiving old/outdated repos in place, and then moving them to an "archived" sub-org, and waiting to see if anyone complained.
Previously no one wanted to outright delete or remove repos because of the risk that someone somewhere was relying on it, and also there was no actual downside to just leaving them there (no cost savings, no imminent danger other than clutter, etc), so resources were never allocated to do it. There was always something more important to work on.
In an org with a higher floor of engineering management, a proactive program for removing unused or outdated repos would absolutely be expected though I think.
This is a continual fight for me. At nearly every company I've had to compromise on using a graveyard repo for packages within a monorepo, even though git has the whole history already.
The problem with history is that you need to know when to look. If you're looking for some old code that you know existed but you don't know exactly what it was, you can't just browse to go and find it.
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.
On the same topic but from a slightly different angle - as SOTA models get more capable, the 'quality' and 'feel' of the experience they provide in each domain is heavily dependent on the reinforcement learning the vendor does for that specific domain. After all, many fields have 100 flavors of "good answers," but the model has to pick one answer.
Benchmarks are not very good at capturing this yet. But it could be the case that DeepSeek v4 Pro is 100% as good as Claude Opus 4.7 at scaffolding a basic Rails app, but absolutely terrible at creating a credible business plan that another businessperson would think is real. That's a made-up example, but you get the point.
The end result will be a lot of people arguing about which model is "better," but "better" depends heavily on the task and how that model was trained to interact with the user for that task. Two users may have very different qualitative experiences using the exact same model, despite the benchmarks.
Creating a nice PPT is actually hard because it requires visual capabilities and so-called "computer use" (really, GUI use) of fiddly proprietary software. The nice thing about the coding case compared to a lot of disparate white-collar work is that it's all plain ASCII text. You can already ask a coding model to create a nice TeX/beamer slideshow (or whatever the Typst-based equivalent is) but whether your boss will be duly impressed by that is anyone's guess.
Tangential, but in our opinion corporate PPTX automation is an unsolved problem, even with Claude for PowerPoint (and it's worse with everything else common out there). Its harness (a) is not tuned very well for corporate use and (b) even if it were, fails to manage the specific business knowledge within each org needed to create effective (i.e. audience tailored) presentations.
2) No
3) Pi
4) No, I think these are token burners and makes it practically impossible to track what's happening from a human side.
5) My own tool: <https://github.com/PAndreew/pi_group> - the main benefit is that it's multiplayer so it functions as a regular chat app in which you can invoke Pi in a group chat. This way the agent can have a shared context, but it still have all the features like generative ui, internal browser etc. that other bolted-on solutions like Claude bots in Slack do not possess.
reply