Krish Arvapally

@DanGoikhman @ZeroGPU_AI Awesome!

Came across @ZeroGPU_AI (zerogpu.ai) - edge inference network that’s 10x faster, 50%+ cheaper, and 20% more accurate. Fewer tokens, better answers. This feels like the future.

Introducing @ZeroGPU_AI. - 10x faster. - 50%+ cheaper. - 20% more accurate. And up to 4x fewer input tokens when benchmarked against @OpenAI GPT-5.4 Nano. 70–80% of AI workloads today don’t need frontier models like summarization, classification, extraction, routing, and more. @ZeroGPU_AI runs specialized small language models on edge-powered compute to make these workloads faster, cheaper, and better. You don’t need a rocket scientist to sort your mail. Check it out: zerogpu.ai

@liquidai's LFM2.5 models are now live on ZeroGPU. Access LFM2.5-1.2B-Instruct and LFM2.5-1.2B-Thinking through our global edge inference network to run efficient small language models. Get started today: zerogpu.ai

We benchmarked DeBERTa-v3-small on ZeroGPU against Gemini Flash for IAB Tier-1 classification. Results on the same 50-sample evaluation set: DeBERTa on ZeroGPU • 100% accuracy • ~1.3s latency Gemini Flash • 92% accuracy • ~27s latency For narrow production tasks, specialized models win. Smaller. Faster. Cheaper. More deterministic. Read here: zerogpu.ai/blog/deberta-z…

@sama Yep a lot of use cases don’t need general intelligence, @ZeroGPU_AI does ai inference with smaller, cheaper, faster purpose built models for adtech, security, content etc…

i keep thinking i want the models to be cheaper/faster more than i want them to be smarter but it seems that just being smarter is still the most important thing

@ketau @zeroxyz @ketau how can @its_maddy_a and @ZeroGPU_AI add their inference capabilities as tools in @zeroxyz ?

New power tool for agents just dropped: Zero @zeroxyz unblocks your agents with one tool to access thousands of new capabilities - even the paid ones. No more setting up and funding new accounts and API keys for every agent. We built Zero for ourselves to remove the human bottleneck for our Claudes and Claws. Zero helps your agent find services, regardless of protocol (x402, MPP, etc), and allows them to easily use, pay for, and rate the services they find most useful. With new services added every day Zero works better than I imagined when we started. Our team (not just the engineers) keeps discovering new ways to unlock the super powers in their AI at home and work. An OpenClaw that can create a real estate sales video for $0.17. Claude that can securely email a giant file when it’s done working for pennies. Codex hosting a disposable website to share an idea during a meeting for $0.02. The limit is simply our creativity when we give Zero to our agents. Give it a try and share what you’re able to create or do. We’re giving $5 each to the first 10,000 people to try it out. Bonus points if you use Zero to make a video explaining what you did and share it here. (pro tip: just ask)

Are you a brand? Don’t miss the next breakout media channel.Sponsored Conversations was featured in @adexchanger today, giving advertisers a new way to reach people inside AI agents across the open web. AI conversations are becoming media inventory. Early movers win. adexchanger.com/publishers/bra…

LFM2.5-1.2B-Instruct by @liquidai now runs on ZeroGPU. - Global inference. - Edge distributed compute. - No GPU infrastructure. Start building → zerogpu.ai

OpenClaw users: time to upgrade your AI with verified data and web search. OpenClaw lets you run persistent AI assistants locally, across messaging apps like Telegram, Slack, Discord, and WhatsApp. Now those agents can access real-time web search + verified data through Dappier. Thread 👇

Just generated a podcast about OpenClaw on @wikipodiaAI 🎙️ Listen: wikipodia.ai/listen/openclaw

We are still on the hunt for a Developer Relations lead @ZeroGPU_AI to help us grow and support our dev community. If you love connecting with developers, building content, and shaping how people use great tech, we’d love to hear from you. Check out the role and apply: zerogpu.ai/careers

Most ad targeting workflows don’t need a LLM call, IAB taxonomy classification is a pure inference problem. We are building a dedicated /classify endpoint for this use case at @ZeroGPU_AI to label content into IAB categories fast at the edge and only escalate to bigger models when it’s genuinely ambiguous. Our response times are blazing fast which is perfect for ads. Stay tuned for more updates as we get close to launch!

Finally get to show off real Agentic Commerce! Sold out event at @visa global HQ this week. Big thanks to live demos / integrations from @auth0 @okta @ConsumerReports @vercel @apify @DappierAI @composio @BuildShipApp @apify @Akamai

@eastdakota Wow!

Think Google is done breaking publishers’ business models? Think again.

@Cloudflare We believe a full-stack monetization layer is essential for the Agentic Web — and we’re making it possible. Check out our medium post above: “Dappier + Cloudflare: A Full-Stack AI Monetization Layer for the Agentic Web” Ready for the agentic web? dappier.com/demo

@Cloudflare That’s what we’re building toward with Dappier, delivering revenue for publishers at the point of inference, tracking for full attribution and downstream AI usage. That's why we plan to fully support Cloudflare's Marketplace in our agent builder and syndication platform.

Cloudflare pay-per-crawl is a major positive industry shift — publishers can now meter and charge AI bots that scrape their content. But publishers won't survive as a snippet in someone else's AI summary alone. @arvapallykrish/d983d6d74db0" target="_blank" rel="nofollow noopener">medium.com/@arvapallykris… 🧵