steve beyatte

Hello everyone!

I'm Chandan, founder of Geekflare, a leading business & tech publishing platform. More recently, I'm building Geekflare AI.

Hey PH fam I launched Lumoar (B2B SaaS startup) three weeks ago and we have already seen over 100 users sign up and lots of positive feedback. It s clear there is a demand and I feel there s potential in this. But now I have hit a wall. Because of my country, international payment platforms like Stripe or PayPal don t provide API access so I can t monetize or implement a paid plan right now. I ve talked to some investors, but they expect initial capital or revenue before they d consider funding. Bootstrapping got me this far, but moving forward without monetization or investment is getting really difficult. If you ve dealt with something like this (or just have thoughts), I d love to hear: What would you do in this situation? Any creative approaches, funding workarounds, or even alternative payment methods I should look into? Appreciate any advice

I'm a founder in the depths of pivot hell. But I have a strong feeling that the important space to build in is memory. For me, the thing that can increase the value we generate from AI 10x from where the models currently are, even without further improvement, is memory and context. Being able to transfer context from one interaction to the other, preferences for how responses should be structured, permission management, sensitive information filtering and redacting, etc. Of course, all these have to be done securely and end-to-end.
I've thought a lot about how to find this wedge. A lot of things I've found just sound cool but probably aren't providing much value to the user. I'm optimizing for a high-value, high-frequency use case. This is the best-case scenario. Especially for a consumer-based solution. I would list some of the solutions that I think can stem from having an end-to-end encrypted memory of all AI interactions (limited to conversations with AI models for now) and a system that does a basic filtering of sensitive information and PIL:

• Smart search: being able to semantically search through old conversation history to find working solutions with pure conversation.

• Prompt injection: Being able to use old verify prompts with a click to ask better questions.

• Conversation summaries: Just as it implies, get summaries of conversations, perhaps of just one chat or multiple. And maybe even across multiple AI tools.

• Project workspaces: Having an organization share a connected memory. With each team member being able to use the other members' context on what they're working on to make sure they're all aligned on the goal and keep a coherent implementation of the tasks.

• Chat with memory: Using a lightweight LLM to converse with your chat history across all AI use and different conversations and sessions.