Notte / Browser Arena vs Together AI Llama 3.3 Fine-Tuning API

“The open benchmark is the ballsiest move here — publishing your full execution traces so anyone can verify your claims is rare in this space. Sub-50ms session spin-up and 47s task completion vs Browser-Use's 113s are meaningful numbers for production agents where latency compounds. SOC 2 already sorted is a big deal for enterprise deals.”

“The primitive here is clean: submit a dataset, get back a LoRA adapter, deploy it — no CUDA drivers, no FSDP config, no sacred Hugging Face trainer incantations. The DX bet is to hide all the distributed training complexity behind a single API call, which is the right call for 80% of fine-tuning use cases. The auto-eval runs are a genuinely useful addition — getting a held-out eval without writing your own harness is the kind of thing that saves a Tuesday afternoon. My one gripe: the 'one-click deployment' language is landing-page speak until I see the actual API surface for versioning and rollback. If that's solid, this is a legitimate skip-the-weekend-script win; if it's a button in a dashboard with no programmatic control, it's half a tool.”

“The benchmark tasks they chose almost certainly favor their architecture — that's how every vendor benchmark works. '79% success' sounds great until you ask what tasks, what websites, and whether those tasks reflect your actual use case. Browser automation reliability degrades fast once you hit sites with aggressive bot detection like LinkedIn or Cloudflare-protected pages.”

“The direct competitor is Modal plus Axolotl, or just calling the OpenAI fine-tuning API — and that comparison is where Together has to win. They do have a credible answer: Llama 3.3 is open-weight and OpenAI won't fine-tune it for you, so if you want this specific model, Together is a real option rather than a convenience wrapper. The scenario where this breaks is at scale: teams with large proprietary datasets and strict data residency requirements will hit contractual blockers before they hit a technical one. The 12-month kill scenario is that Meta ships a hosted fine-tuning offering tied to its own inference cloud, or Groq and Fireworks match this and compete on price, squeezing Together's margin to zero on a commodity service. What would have to be true for me to be wrong: Together builds enough workflow lock-in through evals, versioning, and deployment that switching cost exceeds the price delta.”

“Open benchmarks are how maturing ecosystems establish trust — the same way MLPerf did for model inference. If Browser Arena catches on as the standard, it could do for web agents what SWE-bench did for coding agents: create a common scoreboard that drives genuine competition on real-world capability rather than marketing claims.”

“The thesis here is: within 2-3 years, fine-tuning open-weight models becomes as routine as calling a hosted API today — the infrastructure friction is the only thing stopping most teams from doing it. That's a falsifiable and plausible bet; the trend line is the declining cost of LoRA training on commodity hardware, and Together is early-to-on-time, not late. The second-order effect that matters isn't that teams customize Llama — it's that model customization stops being a specialized MLOps discipline and becomes a product feature anyone can ship, which shifts power away from model providers with closed APIs toward whoever controls the fine-tuning workflow layer. The dependency that has to hold: open-weight models must remain competitive with closed frontier models for the tasks where fine-tuning provides the edge. If GPT-5 or Gemini 2.x make fine-tuning irrelevant by being few-shot-capable enough for every use case, the whole thesis collapses.”

“For anyone trying to automate content research, competitor monitoring, or social listening at scale, reliable browser agents are the missing piece. Notte's hybrid approach — script first, AI fallback — sounds like the right architecture. Looking forward to seeing this mature beyond beta.”

“The buyer is an ML engineer at a mid-size tech company whose team doesn't want to manage GPU clusters — that's a real person with a real budget line. But the moat here is essentially zero: this is compute arbitrage plus a thin API wrapper, and every inference provider with spare H100s can ship the same thing in a quarter. The pricing scales with training compute, which means Together's margin collapses exactly when the customer is getting the most value — high-volume fine-tuning jobs. What would need to change: Together would need to build proprietary eval infrastructure, dataset tooling, or model versioning deep enough that the workflow lock-in survives a 40% price cut from a competitor. Right now it's a good product that isn't a good business.”