Multiple Threads

Localmaxxing : pushing more inference to local models. Over five weeks, I tested how much of my daily work can run on a local 35B model instead of cloud frontier models. The answer : half. Many reasons to use local models : privacy, cost, asset depreciation. But the only one that really matters is latency. I ran a head-to-head benchmark. Qwen 3.6 35B-A3B-4bit on my MacBook Pro M5 vs Claude Opus 4.5 via API. Result : 2.1x faster locally. Mean 2.8s vs 5.8s. The local model isn't smarter. Opus scores ~20% higher on reasoning benchmarks. Local models lag frontier by 3-4 months, and for complex tasks, that gap matters. But for routine agent tasks, it rarely does. If half the work runs 2x faster on my laptop, I'll take that trade every time. My little computer is about to earn its keep. tomtunguz.com/localmaxxing/

Localmaxxing : pushing more inference to local models. Over five weeks, I tested how much of my daily work can run on a local 35B model instead of cloud frontier models. The answer : half. Many reasons to use local models : privacy, cost, asset depreciation. But the only one that really matters is latency. I ran a head-to-head benchmark. Qwen 3.6 35B-A3B-4bit on my MacBook Pro M5 vs Claude Opus 4.5 via API. Result : 2.1x faster locally. Mean 2.8s vs 5.8s. The local model isn't smarter. Opus scores ~20% higher on reasoning benchmarks. Local models lag frontier by 3-4 months, and for complex tasks, that gap matters. But for routine agent tasks, it rarely does. If half the work runs 2x faster on my laptop, I'll take that trade every time. My little computer is about to earn its keep. tomtunguz.com/localmaxxing/

$60k worth of credits live on build something wonderful’s founder stack Unlock partner credits across categories to build faster. ~ ai stack by @OpenAIDevs ~ get an ai domain for 2yrs by @aigrantsindia ~ email stack by @AutoSendEmail ~ media stack by @fal ~ search stack for agents by @ExaAILabs ~ payment stack by @dodopayments ~ video generation stack by @morphic

✉️ Trying @Cloudflare's new Email Sending feature today If you send 1,000,000 emails per month: - Postmark: $1,206/mo - Resend: $650/mo - SendGrid: $600/mo - Cloudflare: $354/mo - Amazon SES: $100/mo So Postmark is now by far the most expensive email provider And SES and Cloudflare are now the cheapest email providers I know my friend @marckohlbrugge is trying out SES now so I'll try Cloudflare and see how it is, SES is cheaper but Marc said it takes a bit more managing, and since I already use so much Cloudflare stuff it's nice to use them for email too With AI especially all of these are just as easy to use and setup in your app/site so economically it makes sense to go for the cheapest, because email is just email, it's all the same and deliverability is good with all of these I think TL;DR email sending has become a commodity!

@levelsio @Cloudflare Very nice stats. When setting up Resend, it actually gives me even the Amazon SES domains...