GM world-changers. Welcome back to yet another edition of the Leaderboard, today we have: a new way to debug at lightspeed, a new trend for AI developers, Amazon's answer to Temu, and of course, the top products of the day. Let's get started.
Bugs running scared
Oopsie: Debug apps with AI and session replays.
Session replay has long been on the radar for the most popular mobile observability tools like Sentry to make it easier to diagnose user-facing problems. New player on the scene Oopsie (available only on React Native and Flutter at the moment) claims to be theĀ onlyĀ AI-powered mobile debugging tool focused on that use case, but I'd be interested to learn whatĀ InstabugĀ makes of that claim since they've been adding similar enhancements this year. In any case, all efforts to make debugging easier are welcome and apps that haven't already made similar integrations should give Oopsie a look.
Ship better AI
Coval:Ā Simulation and evals to ship voice and chat AI agents.
Building voice agents with LLMs once meant battling latency, interruption handling, and endpointing. Now, frameworks like Vapi and Pipecat make it easier to create agents that perform well āmost of the time.ā These advancements have pushed voice AI ahead of human agents in many customer support settings, while enabling 24/7 phone support, healthcare check-ins, automated freight scheduling, and experimental voice UIs. But āmost of the timeā isnāt enough. To achieve consistent reliability, tools for real-time observability, testing, and performance metrics are essential. Coval and Canonical AI are leading efforts here, setting the stage for 2025 to be the year of AI evals and metrics.
Your agents stopped answering and started doing
Your agents call tools, write to databases and move money, so a wrong answer is now a wrong action, which is a considerably worse meeting. Arcjet runs policy before every LLM call, tool call, query and API request: prompt injection, sensitive data, bots, rate limits, spend quotas, and your own rules in Rego on top. It reads the OpenTelemetry you already emit, so switching it on is one environment variable, no code changes and nothing to deploy. Each check also reads the earlier steps in the same workflow, so an agent's tenth action gets judged against the first nine and not on its own. JavaScript, Python and Go SDKs all at 1.0, agent framework guards included, free tier to start on.
Can't beat the original
Amazon Haul: Amazon's answer to Temu
Iāve never shopped on Temuāpartly out of trust concerns and partly because Iām wary of supporting a race to the bottom on prices. But seeing Amazon Haul ābacked by Amazonās A-to-Z Guaranteeā intrigued me enough to give it a shot. Unfortunately, the experience has been disappointing. The app feels cluttered, I canāt easily search for items, see prices, or even add things to my cart (maybe itās overloaded with Black Friday traffic?). Compared to Amazonās usual streamlined interface, this feels chaotic and full of friction. Maybe Gen Z is fine with the busy design, but I just want a quick, seamless shopping experience. For now, I think Iāll stick with the main Amazon platform.
Daily Top Products
