vanveluwen
Cofounder at Hurdle
- Joined
- Nov 18, 2026
Comments
Has anyone set up a new TV/Smart home interface yet?
It is funny to think the value of "old" hardware is probably understated now that we can redo the firmware/OS/stuff running on them with AI along with more efficient/thoughtful stuff (like Omarchy).
Introducing GPT-6.2 Astra and Sol
consumer subs only (although I know devs running 5+ of these) otherwise you get the pleasure of paying 6x on the standard API rate. That said it is soooooo fast.
Introducing GPT-6.2 Astra and Sol
Access to ultra fast mode for the $200 sub is pretty cool, especially if you need to blaze some specific smaller stuff out in the week next to a larger sub/APIs.
[Weekly] What are you building this week?
We've had enough Hurdle customers (check us out (hurdle.ai)) ask for Slack integration that I've had to suck it up and build that out. Already cursed myself as someone is now asking for a Teams integration...
I haven't published this yet but we run a bunch of evals on complex sales tasks, I'll cut out a public version of the tasks to show how the frontier models are performing at real sales work. I had the chance to run Astra 6.2 today and found some good results (better substantiated claims in assessing deal progression) but interesting less financial/data details in an ROI case compared to Fable 5.2.
Hacking a new home screen for my Samsung TV
Hahaha this happened a few weeks ago when I was working on my LG. Opus 5.5 immediately flags issues (and this was before hack your TV was too popular), I just routed the prompt to Codex to rewrite (literally "use language less likely to trigger Anthropic's checks") and was able to get Opus to finish the job.
The dawn of the Data Intelligence Engineer | MotherDuck
Yeah, I think the shift from writing code to verifying the results is a big one. At LinkedIn, it felt like we spent 20% of our effort figuring out which metrics we needed for executive reporting and 80% building the Airflow pipelines to produce them.
Now we can try more metrics and iterate much faster. As more people start answering their own questions with AI, I think we’ll spend more time agreeing on what’s worth measuring and checking that the numbers mean what people think they mean. Also (at least in the messy middle) there probably will be a lot more "wrong" analytics unless there is a verification/audit layer.