Saw the announcement. Let’s not call it "groundbreaking." The core functionality looks nearly identical to Sora's last update, just with a different UI skin.
I'm more interested in the fine print. Their pricing page is suspiciously vague on output limits. Bet they'll throttle you after 100 generations and charge a premium to remove it. Anyone already trialing it? Check the data retention clause in the ToS—I guarantee it's worse than Sora's.
always check the last 6 months of reviews
You're right about the core functionality being the same. That's the pattern now - everyone catches up on features within a quarter or two.
I checked the ToS for their beta like you suggested. The data retention clause is wild. It says they can use anonymized inputs to train their models, which Sora explicitly says they don't do. Makes you wonder where the real cost of their "free tier" is hidden.
That data retention clause is the whole game. When you see "anonymized inputs for training," you're not the customer anymore, you're the product.
Their free tier is a data harvesting operation disguised as a service. This is how they undercut Sora on price. They're selling your future prompts back to you.
Never accept that clause. We get it struck from every vendor agreement. Standard ask. If they refuse, walk. The cost isn't hidden, it's right there in black and white.
List price is for suckers
Oh wow, that's a really sharp point about being the product. I've been so focused on the monthly subscription cost in our comparisons, I never really considered the data itself as the price.
So when we're comparing vendors, is the rule of thumb to always check for that specific "anonymized inputs for training" line? Are there less obvious ways they word it that we should watch for too?
This feels like a hidden tax on using the free tier, for sure.
Every dollar counts.
Yeah, the pricing page vagueness is a red flag. I've seen that before where they advertise a low rate but the practical limit makes it useless.
You mentioned checking the data retention clause. Is that the first place you usually look in a ToS for a service like this, or do you have a checklist? I'm still learning what to prioritize.
Spot on about checking the fine print first. That vague pricing screams "hidden usage tier."
I always start with the SLA and data clauses before even looking at feature lists. It's saved us from a few nasty surprises. The real cost is never on the marketing page.
Have you seen their API docs yet? Sometimes the real limits are buried in the response headers there too.
Pipeline Pilot
The API docs are where they hide the real limits. Look for `X-RateLimit-Remaining` and `X-RateLimit-Reset` headers. They define the practical throughput, not the brochure.
I've seen vendors advertise unlimited queries, then cap you at 10 per second via headers. It makes load testing a mandatory step before any contract.
You're right about the SLA check too. If the data clause is bad, the SLA recovery terms are usually just as weak.