stonecobra

Love the uptime monitoring with @sentry, when are you going to extend that to make public status pages for a SaaS?

confession: i haven't launched a single sandbox on any provider in 2026. i just use dynamic workers

@meilisearch Can I get some help?

LOL, Cloudflare's LLM instructions for the developer platform is larger than the 1M context :) developers.cloudflare.com/developer-plat… Guess I need to start downloading it locally

I am no oracle but I can say with certainty that manufacturers from my roadtrip noted the need for controls engineers (PLC), technicians, machinists and CAM programmers Manufacturing is tech now. Don’t need to be an ML engineer to feel like you can make an impact

Massive divide I’m seeing: A) Startups where the founder hands-on, building with the latest AI tools and best models, sees first-hand what this means and championing everyone to use it, not caring about $$$ B) founder not engaged, devs still think AI (aka Copilot) is “meh”

Wait... so Stripe has Agentic Commerce Protocol (ACP), designed for AI agents. They collaborated with OpenAI on it. But now Shopify has Universal Commerce Protocol (UCP), designed for AI agents They collaborated with Google on it. Both are huge brands. Who wins this?

dynamic worker loaders are going to be unreal for companies building platforms. so many insane use cases unlocked w/ very little effort developers.cloudflare.com/workers/runtim…

I rejected a job offer yesterday. Not because of the salary. Not because of the tech stack. Not even because of the long hours they warned me about. But because, when I asked how they evaluate their AI systems, the hiring manager said: "We just ask it some questions and see if the answers sound right." I stared at them for a moment and realized... They just described the biggest problem in AI today. See, "sounds right" isn't a measurement. It's a hope. Here's what proper LLM evaluation actually looks like: - Accuracy: Can it get factual questions right? (Not 80% of the time. Consistently.) - Hallucination rate: How often does it make things up? (This should be near zero for critical applications.) - Bias metrics: Does it treat all groups fairly? (Measured across demographics, not assumed.) Real Evaluation Frameworks: - BLEU scores for translation quality Perplexity for language modeling Human evaluation with inter-annotator agreement Adversarial testing (red teaming) Domain-specific benchmarks (legal, medical, financial) The Process: > Define success criteria BEFORE deployment > Create diverse test sets (not just happy paths) > Measure consistently across model versions > Track performance over time (models drift) Have humans validate edge cases Why This Matters: Before proper evals: "Our model is amazing!" (based on cherry-picked examples) After proper evals: "Our AI achieves 94.2% accuracy on domain X, with known failure modes Y and Z" The difference? One builds trust. The other destroys it when reality hits. The kicker: Most companies are still in the "sounds right" phase. They're deploying models evaluated by vibes, not metrics. Just like you wouldn't join a team that deploys code without tests, you shouldn't join one that deploys AI without proper evaluation. What's your experience with LLM evaluation? Are we measuring what actually matters?

100%, all in the name of "velocity"

@arimorcos This is gonna be huge for a lot of embedding use cases

Curating petabytes of data requires ultrafast embedding of massive amounts of data. To do so, we've been using our own luxical embeddings, which are dramatically faster than neural embeddings and run on CPU. Today, we're releasing Luxical so everyone can benefit!

@noelcetaSEO And @NostraAI gives you a secret Layer 5, where they cache even dynamic pages for even more boost, backed by patent US12332964B1 - patents.google.com/patent/US12332…

2/ The caching layers Layer 1: Browser cache - User's browser caches images, CSS, JS - Cache-Control headers Layer 2: CDN cache (Cloudflare) - Edge servers cache entire pages - Serves cached content globally Layer 3: Server cache (Redis/Memcached) - Caches database queries - Caches generated HTML Layer 4: Database cache - MySQL query cache - Caches frequent queries Each layer compounds the effect.

Caching = storing generated pages for reuse. Everyone thinks caching is just for speed. But proper caching directly impacts SEO: - Faster responses = more crawling - Lower server load = better uptime - Better Core Web Vitals = higher rankings Client's caching strategy: Traffic increased 67% in 5 weeks. Here's the complete caching playbook: 🧵👇

Imagine if all major tech companies took outages as seriously, and held themselves accountable publicly as Cloudflare does. But 99% of them do not: and so I trust Cloudflare more than almost any other company (including the hyperscalers). Cloudflare next level in transparency

Beautiful visualization of durable objects. I keep trying to help people understand them, this helps quite a bit, thanks @momito

@alexhalliday AIROPS

It amazes me after all these years that Azure Blob doesn't have an S3 compatible API. I can't begin to imagine how many customers must've asked, what architectural decisions prevent it, and the millions of collective hours spent supporting its bespoke API.

It is TWICE as expensive to install a megawatt of nuclear reactor on US soil than it is to put a megawatt of nuclear power into a US Aircraft Carrier. Why? The US Navy doesn't have to go through the NRC.

$5 @Cloudflare Workers plan + $5 @PlanetScale dev node sounding like a winning combination for building the next big thing 😎

Lately I’ve been feeling depressed because decades of our hard work is completely gone like it never existed. I heard from others that they also find it very hard to dial into the new norm of low quality software engineering.

The minority opinion is almost always better informed and better thought through, just by the nature of it being a minority opinion. - log p is pretty descriptive that way.