smadam9

I built a personal tool at https://shortlist.addslift.com that is scanning jobs regularly. I'd love to integrate in your dataset. The CSV is a great start, but if you ever choose to create an API for it, let me know.

Love the idea. Don't love that you require my email to test things out.

What are the main differences to Glean? My company is evaluating Glean and I feel like Airbyte is a strong alternative (at least for some use cases).

How does Airbyte handle data authorization?

Feel free to get in touch! I'm happy to collaborate to make something greater

I did something similar with https://shortlist.addslift.com/

I've since heard from so, so many people that they aren't getting any bites with traditional job applications, so I've paused active development on it.

My motivations were very similar to yours. I have the personal opinion that there is so much more that we have in our heads that we don't put in our resume/CV and that's the approach I took. You need to come in with a sort of AI biography/career highlight reel on yourself to get the best results. I provided users a template to start from. Yes, your resume is a source of the bio, but there is so much more that job hunters should be drawing from.

That AI biography was the result of another project I was working on.

How do you link the cover letter/resume to a job? Does the user need to come in with a job description in mind? (I didn't try your tool out yet!)

I understand your frustrations. I feel your approach works well for a narrower profile. One person, stable context and fewer files.

I've seen many users that need a wider breadth of memory across more topics, where structure and organization of that memory plays a big part in the LLM's performance.

My response to that was a local system that I ended up turning into Sig <https://sig-ai.app/>

It has some overlap to how you've approached it, but differs in other obvious ways.

Having said all that, I'm just highlight another use case for memory. I think your approach is a very valid approach for a lot of people. I appreciate the simplicity and lack of lock-in.

I've done something similar locally.

I've used a lot of exercise science data that I've pulled from various sources and "trained" and agent to construct my workout program based on goals, current health, etc., all also pulled in from various data sources.

My focus lately has been mobility, so guys like Vernon Griffith have been some of main training material for my local agent.

I would love if this could be monetized for the creators/knowledgeable folk. So that e.g. Vernon could get a commission when an agent uses his knowledge.

Anyway, great that you've published this! I hope you extend it further. ...mine is currently expanded toward cardio and overall health. I love it :)

The screenshot feedback is fair. Updating it. The capture response should feel like "filed. here's where." not a report back at you. Working on it.

Team Sync isn't a group chat or a message sender. It's more like an approval step: you review what you've captured privately, decide what's actually worth sharing with your team, and publish that specific text to a shared knowledge base. Right now "publishing" means an abstracted git flow that pushes updates to a central Git repo on Github. Nothing goes to the team without you explicitly choosing it. The name could be clearer — that's useful feedback, thanks.

On integrations: Slack imports work today (you can pull exports into your context). Teams Meeting Summary is an interesting one. Right now you'd paste it in and Sig routes it. The summary gives you the factual scaffolding, then you add your layer on top. That's the part the transcript can't give you.

A comment here mentioned "Everyone is writing. Nobody is reading." but I think the friction starts even further upstream.

I've been building a native Mac app "Sig" around this idea: capture has to come from you. You sit down after a meeting and articulate what happened — what was decided, who committed to what, what you actually think. That articulation is the work. The AI routes it into files. If you skip that step and scrape transcripts instead, no promotion workflow saves you. You're just contributing to the garbage in, garbage out idea.

That is Sig <https://news.ycombinator.com/item?id=47901737>

You talk about your work. Sig builds your KB and helps you contribute to your team's.

Most of what actually happens at work never gets written down. Meetings end, decisions get made, people commit to things verbally. All that lives in someone's head until it doesn't. The tools that exist (Notion, Confluence, Glean, etc.) can only search what's already been formally documented. That's usually not the useful stuff.

Sig is a macOS app for non-technical knowledge workers. You talk about your work: a meeting you just left, a decision that got made, tension you noticed. Sig routes it into two knowledge bases:

Your personal KB — private, on your machine, organized across sessions

Your team KB — only gets what you explicitly review and approve

One thing it isn't: automated note-taking. You still have to sit down and describe what happened. That's intentional. The interpretation comes from you, not a transcript.

Everything is plain text on disk. No server. Works with Claude, ChatGPT, Copilot, Gemini or whatever you already use.

Currently in private beta. Public release very soon.

Yes, the .md's are in their own repo, locally. The entire UI is a layer on top of that repo. The UI has some underlying mechanisms that abstract the git operations away from the user, but that doesn't stop a power user from jumping in the shell and accessing the repo directly.

The "magic" starts when Sig contributes to another, remote repo - a central knowledge base that all teammates' local Sig can pull from, and contribute toward.

You beat me to it by a day! But well done Luca. The tool looks excellent and I'm trying it out now.

I'm building Sig <https://github.com/adamjramirez/sig-releases> and the architecture overlap is obvious: macOS, plain markdown, git-versioned, designed as context for AI agents.

The difference is where in the workflow we start. Tolaria seems to excel at organizing knowledge that already exists. Sig is trying to solve what happens before that - how to get the knowledge out of your head and into files in the first place. Most of what actually determines the quality of your AI output was never written down: the decision made in the last five minutes of a meeting, the verbal commitment with no follow-up, your actual read on what a conversation meant (not the surface version).

Sig's capture is two layers: 1) factual record first, 2) your personal interpretation on top. Both stored as markdown on your machine. When you're ready to share to a team knowledge base/open brain, it's an explicit decision to do so and opt-in — private by default, team-readable only when you choose.

I was curious about how the creator economy actually works, not the influencer marketing pitch decks, but the actual structure.

- Who mentions who in their videos?

- Which brands keep showing up across the same creators?

So I built a pipeline that analyzes YouTube videos at scale, extracts every person mentioned and every brand sponsor, and builds a relationship graph from it. No manual data entry. Everything is extracted from what creators actually say and show.

Current data: 32K creators, 55K relationship edges, 2,700 brands. Started deep in fitness and gaming YouTube.

Things to try out:

- Search "Gymshark". See their sponsored creators, which categories they target, and which brands compete for the same roster

- Search "MrBeast". See his connection network

- Click a category tab on the homepage to filter both creators and brands

The hardest technical problem was entity resolution. The same person gets mentioned as "Jeff Nippard", "Jeff", "Nippard" across thousands of videos. Getting that right is what makes the graph useful.

Stack: FastAPI, Next.js, PostgreSQL, Fly.io. Gemini 2.0 Flash for structured extraction.

No signup, no paywall.

There was a post a few weeks ago for a site dedicated specifically to this idea: https://getstartupbusinessideas.com/

https://tangerinefeed.net/

It is lacking on features at the moment, but if you are willing to do some manual filtering, you have more or less first access to a lot of jobs.

Edit: Giving credit where due (original post) - https://news.ycombinator.com/item?id=38177286

I'm curious to learn how recruiters distinguish self-written vs. AI-written cover letters.

"Put your kids on the blockchain", as in setting them up with a wallet from early on so that it can be accessed later?

Location: Dallas, TX

Remote: Yes

Willing to relocate: No

Technologies: Python (Django), AWS (EC2, ECS, ALB, Lambda, DynamoDB, S3, CloudFormation, IAM, SQS, SNS, StepFunction & many more), Docker, Docker Compose, Redis, ElasticSearch, Postgres, MongoDB, GraphQL, Bash, Flutter, Riverpod

Résumé/CV: Please message or email me

Email: mradamjramirez at gmail.com

---

With around 12 years in tech coding & managing, I'm your hands-on tech leader, specializing in infrastructure and backend development. I've steered teams through startup hustle, growth vibes, and even the occasional merger chaos. Actively involved in crafting systems from the architecture to the coding. Open to management or hands-on coding roles (backend, DevOps, Infra, Platform, etc.)

0.25€ per kWh

The same principle applies with the airline. If you really don't want to deal with the hassle of having your bags lost, don't fly. Of course, that takes the whole mentality to an extreme. It's not often, even with a good privacy policy, or lost baggage policy, that it offers any true protection. It is more for peace of mind.

If you don't want your financial information shared with anyone, don't put it on the internet. Even with a strong privacy policy, do think anyone can guarantee that your data will always remain safe?

Where did you find the dev to help you out?

Lovely. Exactly the sort of thing people into "economical order quantity" ordering are into.

Plans to expand?

Graylog2 (http://graylog2.org/) - Log management & analytics in browser

Torch (for hosted Graylog https://www.torch.sh/)

Is it not frustration, as opposed to anger?

I think the "this doesn't belong on the first page" (or such) applies to a lot more than just the "$x on my last eBook" stories.

I think I've seen that 'O' logo somewhere...?

Clickable faces & a bio on each, would have been awesome. Otherwise, really neat.

I'm not sure there is a one-fits-all answer here - specifically regarding disk space & bandwidth.

Are you saving HD video files? Then you will probably need a data store connected to your web server.

Are you just allowing people to post text, small images and such? In that case, 1 server could easily cut it.

With proper software configuration, and a decent hardware config, you can go quite a long way with 1 machine.

Can you provide any details? Maybe someone can help out with your specific question, since the general question you've asked is quite difficult to give a proper answer to.

How is that similar? I didn't see any opportunity to compare websites on criticue.