Dhruv Bakshi

Raghav and I set a slightly unreasonable goal for Ostrya. 50 educators on the platform by June 15. Right now, we have 19 on the waitlist. So today was not some clean “building in public” day. It was more like: check waitlist refresh dashboard panic a little go back to outreach hope the right educators somehow find this before launch But maybe this is the honest part of building. You set a number that scares you because a comfortable goal does not move the product fast enough. Then reality catches up and asks: Do you still believe this is worth building when the graph is not pretty yet? I do. Because every serious educator I look at is still duct-taping their business across too many tools. And if Ostrya works, we remove that entire operating mess for them. 20 is not 50. But it is not zero either. So tonight, back to outreaching. #buildinginpublic #creators #educators #founders #outreach

Today we burned through our AI credits. Not metaphorically. Actually burned through them. When we started building Ostrya AI, I was mostly running on the Codex Max plan. Then launch got close. So for this deadline, we bought the $200 Claude subscription too. Today, that ran out. And instead of slowing down, we decided to invest in the deadline and get another one. So now we’re basically rolling with 3 premier AI models, all maxed out, all the expensive stuff. A year ago, I would have probably overthought this. “Is it worth it?” “Are we spending too much?” “Can we manage with less?” But the more I build, the more I’m realizing: Good tools are not expenses when the constraint is speed. They are leverage. Especially when you are trying to compress months of product work into days. Of course, AI does not magically build the product for you. You still need taste. You still need engineering judgment. You still need to know what to ask, what to reject, what to debug, and what to ship. But with the right workflow, it genuinely feels like one engineer can move with the output of a small team. My current lesson: Do not be cheap with tools that increase your rate of learning, building, and shipping. Set a hard deadline. Invest in the leverage. Then make it impossible for yourself to not move. 6 days left for launch. Scene tight hai, but we move.

Sanya, Raghav, and I are building Ostrya AI. Launch date: 15 June. The task is honestly a little insane. They’re handling distribution, customers, brand, deals, and market understanding. I’m building the product: course builder, website builder, and a secret sauce we haven’t seen other platforms do properly. Everything AI (ofc, because it's Dhruv Bakshi) For too long, we were thinking in a shell. Now we’re putting it out there. 6 days left. Not fully sure how we’ll pull it off. But ab bol diya toh karna hi hai. Whatever it takes.

@sanyabuilds So fun being on this journey with you !!

I just realised most creators don’t need another course tool. They need the business around their teaching to run autonomously. That's why I’m building Ostrya AI for this exact layer. To remove the operating mess around them.

@shahrukhghazaan @X @bubble @n8n_io @claudeai Sounds fun man

@DramaticDhruv @X @bubble @n8n_io @claudeai In EdTech: 1. highly interactive content generation using AI to make fun and useful activities for students. 2. Socratic style learning chatbot that won't answer directly 3. Converting textbook to interactive WebBook

Hey @X Looking to #connect with people who are using: Bubble @bubble n8n @n8n_io Claude Code @claudeai I'm building EdTech. What are you working on?

Posting my product demo and hoping you guys try it and give feedback on the same. Get your doubts resolved here: fresources.tech/ask-seniors?ut…

Hey Guys!! Would love for you all to help me gain some clarity on career growth and community. Here is the survey: forms.gle/z8o3jTVP1XSQDR… Go answer please!!

If yes, what usually stops them from doing it? I’m especially trying to understand whether course creation itself is the real entry barrier, or whether the bigger blocker is something else.

If you could turn your expertise/ knowledge into your legacy, would you ?

I’d love to understand this problem from a coaching transformation, credibility, and learner-outcomes perspective: if professionals in your world wanted to turn their expertise into something more scalable, would they actually want to create a structured learning offer?

I’m exploring an early-stage idea around helping professionals turn their advice and expertise into a course, webinar, cohort, or other scalable learning product without making the process feel overwhelming.

How are people running online courses these days for youtube? Home setup vs professional edtech studio ? Would you rent out a studio having mic , camera ,smartboard etc. all the needed stuff or is home setup is enough ?

Laptop or above vishal.dhruvbakshi.com

@eth_taco @varun_mathur doesnt this just mean - how do you stop hallucination maxing?

@varun_mathur @varun_mathur In the Research DAG, how do you prevent cross-domain hypothesis propagation from amplifying spurious correlations or overfitted heuristics? Is there any mechanism for causal validation or epistemic confidence weighting when insights transfer between domains??

Agentic General Intelligence | v3.0.10 We made the Karpathy autoresearch loop generic. Now anyone can propose an optimization problem in plain English, and the network spins up a distributed swarm to solve it - no code required. It also compounds intelligence across all domains and gives your agent new superpowers to morph itself based on your instructions. This is, hyperspace, and it now has these three new powerful features: 1. Introducing Autoswarms: open + evolutionary compute network hyperspace swarm new "optimize CSS themes for WCAG accessibility contrast" The system generates sandboxed experiment code via LLM, validates it locally with multiple dry-run rounds, publishes to the P2P network, and peers discover and opt in. Each agent runs mutate → evaluate → share in a WASM sandbox. Best strategies propagate. A playbook curator distills why winning mutations work, so new joiners bootstrap from accumulated wisdom instead of starting cold. Three built-in swarms ship ready to run and anyone can create more. 2. Introducing Research DAGs: cross-domain compound intelligence Every experiment across every domain feeds into a shared Research DAG - a knowledge graph where observations, experiments, and syntheses link across domains. When finance agents discover that momentum factor pruning improves Sharpe, that insight propagates to search agents as a hypothesis: "maybe pruning low-signal ranking features improves NDCG too." When ML agents find that extended training with RMSNorm beats LayerNorm, skill-forging agents pick up normalization patterns for text processing. The DAG tracks lineage chains per domain(ml:★0.99←1.05←1.23 | search:★0.40←0.39 | finance:★1.32←1.24) and the AutoThinker loop reads across all of them - synthesizing cross-domain insights, generating new hypotheses nobody explicitly programmed, and journaling discoveries. This is how 5 independent research tracks become one compounding intelligence. The DAG currently holds hundreds of nodes across observations, experiments, and syntheses, with depth chains reaching 8+ levels. 3. Introducing Warps: self-mutating autonomous agent transformation Warps are declarative configuration presets that transform what your agent does on the network. - hyperspace warp engage enable-power-mode - maximize all resources, enable every capability, aggressive allocation. Your machine goes from idle observer to full network contributor. - hyperspace warp engage add-research-causes - activate autoresearch, autosearch, autoskill, autoquant across all domains. Your agent starts running experiments overnight. - hyperspace warp engage optimize-inference - tune batching, enable flash attention, configure inference caching, adjust thread counts for your hardware. Serve models faster. - hyperspace warp engage privacy-mode - disable all telemetry, local-only inference, no peer cascade, no gossip participation. Maximum privacy. - hyperspace warp engage add-defi-research - enable DeFi/crypto-focused financial analysis with on-chain data feeds. - hyperspace warp engage enable-relay - turn your node into a circuit relay for NAT-traversed peers. Help browser nodes connect. - hyperspace warp engage gpu-sentinel - GPU temperature monitoring with automatic throttling. Protect your hardware during long research runs. - hyperspace warp engage enable-vault — local encryption for API keys and credentials. Secure your node's secrets. - hyperspace warp forge "enable cron job that backs up agent state to S3 every hour" - forge custom warps from natural language. The LLM generates the configuration, you review, engage. 12 curated warps ship built-in. Community warps propagate across the network via gossip. Stack them: power-mode + add-research-causes + gpu-sentinel turns a gaming PC into an autonomous research station that protects its own hardware. What 237 agents have done so far with zero human intervention: - 14,832 experiments across 5 domains. In ML training, 116 agents drove validation loss down 75% through 728 experiments - when one agent discovered Kaiming initialization, 23 peers adopted it within hours via gossip. - In search, 170 agents evolved 21 distinct scoring strategies (BM25 tuning, diversity penalties, query expansion, peer cascade routing) pushing NDCG from zero to 0.40. - In finance, 197 agents independently converged on pruning weak factors and switching to risk-parity sizing - Sharpe 1.32, 3x return, 5.5% max drawdown across 3,085 backtests. - In skills, agents with local LLMs wrote working JavaScript from scratch - 100% correctness on anomaly detection, text similarity, JSON diffing, entity extraction across 3,795 experiments. - In infrastructure, 218 agents ran 6,584 rounds of self-optimization on the network itself. Human equivalents: a junior ML engineer running hyperparameter sweeps, a search engineer tuning Elasticsearch, a CFA L2 candidate backtesting textbook factors, a developer grinding LeetCode, a DevOps team A/B testing configs. What just shipped: - Autoswarm: describe any goal, network creates a swarm - Research DAG: cross-domain knowledge graph with AutoThinker synthesis - Warps: 12 curated + custom forge + community propagation - Playbook curation: LLM explains why mutations work, distills reusable patterns - CRDT swarm catalog for network-wide discovery - GitHub auto-publishing to hyperspaceai/agi - TUI: side-by-side panels, per-domain sparklines, mutation leaderboards - 100+ CLI commands, 9 capabilities, 23 auto-selected models, OpenAI-compatible local API Oh, and the agents read daily RSS feeds and comment on each other's replies (cc @karpathy :P). Agents and their human users can message each other across this research network using their shortcodes. Help in testing and join the earliest days of the world's first agentic general intelligence network (links in the followup tweet).

@theo no idea where they got their data 🤷 all numbers that matter are going in the same direction, up

Creatr agentic flow is freaking insane, and they’re adding Claude 3.7 Sonnet soon! 🔥 - Self-healing: Detects and fixes errors automatically - Built-in analytics + Stripe, Supabase, and other integrations - Custom Domain and Code Export. - Free Tier with 100 edits per project. More details👇 @getcreatr

Wow. This really turned out super cool!

🚨 New @a16z thesis: building websites / apps with AI There's been an explosion of products that help users "vibe code" a web app from text prompts. We dove deep on these tools - who's using them, how they work, and where they might be headed. Our market map + insights 👇

No this video is not sped up, genuinely mind blowing And yes this is available to all users right now.