OptFolio is live on Product Hunt today ๐ฎ๐ณ
Portfolio score from your holdings, momentum and value screens, and optional AI briefings on your holdings, all built for Indian investors.
We'd love your support and feedback: https://t.co/xRA77GNAz2
@aaliya_va Checking numbers is the most underrated use of agents. One wrong stat can wreck a whole thread's credibility, and having sources pulled before you even open Slack is a nice workflow.
@CNBCTV18Live Good move for retail investors. Over 15-20 years, the expense ratio gap between regular and direct plans adds up to a serious chunk of returns. Hopefully distributors also become clearer about what they charge.
@swyx@antigravity@saranormous Curious whether it holds up on long agentic tasks and not just the benchmark. Tool-calling reliability over 50+ steps is where most models still fall apart. Looking forward to your review.
@bindureddy The giveaway for me is that nobody says 'not sure' anymore. Every take shows up perfectly balanced and polished. Real people hedge, ramble and get things wrong in public.
OptFolio is live on Product Hunt today ๐ฎ๐ณ
Portfolio score from your holdings, momentum and value screens, and optional AI briefings on your holdings, all built for Indian investors.
We'd love your support and feedback: https://t.co/xRA77GNAz2
@cmicpace Thanks Colin! That's exactly the gap we kept running into. Most tools treat Indian stocks as an afterthought, so we built OptFolio around NSE/BSE from day one.
@aaliya_va Stairs are where most delivery robots fall apart. Uneven steps, changing angles, and the load shifting all at once. Curious how they handle it when the box isn't balanced.
@aaliya_va@Templafy Brand context is the real gap. Generic AI decks are easy now. Decks that actually follow a company's templates and tone are where agents earn their keep.
@jerryjliu0 About 1c a page for strong table parsing is a good deal. Layout and charts are still the part that breaks most pipelines on financial reports, so pairing it with a dedicated parser makes sense.
@emollick Moving to closed benchmarks makes sense once open test sets start leaking into training data. Curious whether the gap holds on real agent workflows though. Plenty of teams pick open models for cost and control, not leaderboard rank.
@dmuthuk Sensex down 700 today and this is exactly the reminder people need. The hard part isn't knowing what to buy, it's not hitting sell at the bottom. A paper loss stays on paper unless you make it real.
@aaliya_va The valuation is the headline, but I'm more curious about gross margins. For a frontier lab, compute cost is the line that decides how this trades.
@JustAnotherPM The biggest gap I see when PMs work on agents is evals. Defining what 'done right' looks like for a multi-step task matters more than which framework you pick.
@_vmlops Good to see it focus on running agents in real workflows, not just prompting. Curious how much it covers tool permissions and reviewing agent-written PRs.
GitHub just dropped a certification for Agentic AI Developers
Learn how to build, manage, and orchestrate AI agents across the software development lifecycle.
If you're working with AI agents, this one's worth checking out ๐