What vector DB are you using? What is the data structure that you're vectorizing? What is your chunk size? Have you implemented memory? What prompt or technique are you using (ReACT, CoT, few-shot, etc)? Are you only using vector DBs? Do you use sequential chains? Does it need tools? Depending on your data, business case and what output you expect from the agent, there is no one-size-fits-them-all.
The solution to this problem is:
- ban centralized exchanges
- enforce stable coins to have their own coin-to-fiat exchange
- allow only decentralized exchanges that can operate only if they implement some interoperability standards like dex-to-dex bridges
- allow coins on decentralized exchanges only if they implement interoperability standards like blockchain-to-blockchain bridges
Now everybody can safely swap and transfer coins on whatever dex. If they want to cash out, swap again for usdt, then go to usdt official exchange and get your $.
Love the idea! But if i understand correctly, a developer like myself can use your React components to build a workflow that runs on your infrastructure. Neat! But i'm pretty paranoid and i don't want some unknown service provider to peek into my data. Not saying you're doing it, not at all, but i'd rather deploy n8n somewhere myself. Personally I would love to see a React component library like yourself, but could point to my service endpoint. Sure, individual developers will use your work for free, but startups could pay you to take care of their infra
I'm a very easy going person. I find friends everywhere I go. I have also a strong personality and I try to talk about interesting topics. People seem to like listening to me. HN loads me with intell nontheless. Wine is one of the secrets :) I wasn't like this before, but age taught me that others need me as much as I need them. We're human, cheers
I've been helping law enforcement to harvest torrent seeders for the last 10 years. LT is on my list. If there is someone in the world who owns one of the biggest torrent databases, probably it's me
PatternFly is a UI lib with premade components that you can use to build your app with. Storybook is used to showcase your app's UI and people often use it to showcase their own design systems
Think strategically. You want to improve your current product with new features that will bring more revenue, but reality is that you could make even more if you rewrite it. So do it in parallel, plan accordingly and invest wisely. Don't rush, just start and make little steps often. Eventually you will surprise yourself with how much progress you made. The easy path is to get comfortable now but this might expose you in the future.
I've been into software engineering for 8+ years. As any coder i aspired to a faang job. Blew a good chance in my beginnings when competition was not this crazy. Then years passed and later found myself considering how a faang life would've been. I never really managed to discipline myself into faang interview prep. But today i have tons of startup experience, know some good tech, built some products, my passion is strong, and my only reality is that i will eventually succeed with my life plans. We all chase the money, and landing a faang job is a certainty of getting it, but you could regard faang as a false objective. Instead focus on learning how to deal with your life, with your craft, be a little cunning and money has a good chance to follow. It takes time but who cares, enjoy the ride
Understand the company, the product, the business model and the vision. As a lead it's expected from you to be autonomous and take good decisions. I always used my sincere common sense to find answers to these problems and often i took my time to research, ask others, overall i tried to understand the domain as best i could while ignoring the impostor fear, and then things start falling into place. If you like to code, you know how code should be written. If you respect yourself, you know how to handle others. If you ever had your own company, you would know how to treat others' company as well. Don't hire from a dry feeling to step up to your lead role. Hire because you have arguments, hire because the company needs investment in fresh blood everyone and make it fun for everyone
Most probably they gathered feedback and now are busy planning, prioritizing, coding with respect to some new architecture / vision. Also they take their time and don't want to rush. All the planet is hooked and a smart product manager will prefer to dog-feed small improvements over time.
I do recall being impressed by watson marketing at that time. What turned me off was that some of their cool examples were closed source and the whole developer platform felt buggy or under-documentated. Eventually watson faded away as i never saw other people using it. Now it feels like the whole purpose of watson was to make quick bucks to some ibm corporate dudes. Btw what else is ibm doing nowadays, other than riding its own brand and becoming increasingly irrelevant? When i hear ibm i associate it with old people who lost their craft in enterprise nonsense
Yeah one could publish a browser extension that hides these in-app popups, and could even be diligent over time to provide frequent updates to improve identification of such popus that are constantly diversified and need manual maintainer labor to get them all blocked. It would be cool, but i won't pay and anyway broswer extensions don't work on mobile where i spend 50% of my screen time. An alternative is to implement a vpn that proxies the OS's ssl traffic and removes popups from html content, but it has to have system trusted certificates and will break apps with certificate pinning anyway. Your only tool is to make your legislators remove popups the same way they introduced them
Unfit team, unfit org processes, unfit infrastructure, unfit codebase, unfit business model. They don't start happening all at once, but it takes just one of these to corrupt the others in time. Theoretically a good team could have the power to throw everything away and start from scratch, but very few teams can be proud of such feat. Without the right people, incentives, mindset it's almost impossible, because no one has the guts and energy to take responsibility if there is no hope ahead. And passion, it's so important, but very few teams are proud of having passionate people, regardless of age
Same. Copilot can get a bit useful after it understands your intent and starts hinting code from your codebase. Or sometimes i completely forget basic stuff like core functions, algos, patterns, and copilot jumps to help. ChatGPT excels at NLP, parsing or data transformations, but usually i use it for exploring general ideas and not necessarily for coding
Explored the idea of starting a tech coop myself. Haven't really went into too much theory because it felt off. Already established tech coops are closed, have a steady income, started as a group of close friends and they most probably just profit from calling themselves a coop. In reality, you will fail hard and fast trying to find the right people who will accept working on their time to share income with some hippy dude sketching out some figma designs when he happens to be awake. I'd better start a company and learn how to add a profit margin to other people's work
Get a job, finish CS, learn product/marketing/cloud architecture. In any order but the point is you have to remain competitive while others fall behind. Sticking to HN over the years will teach you good lessons btw
You need a way to expose your code to others. Use chatgpt, hire a mentor or reviewer, open source your code, whatever it takes. Code reviews are crucial