Hacker News: gsandahl

New comment by gsandahl in "Show HN: AI Roundtable – Let 200 models debate your question"

gsandahl — Wed, 25 Mar 2026 08:18:54 +0000

Agree, this is where llms can uncover new perspectives!

New comment by gsandahl in "Show HN: AI Roundtable – Let 200 models debate your question"

gsandahl — Tue, 24 Mar 2026 21:16:50 +0000

Oh lord, imagine asking ”serious” questions

https://opper.ai/ai-roundtable/questions/you-are-standing-in...

New comment by gsandahl in "Benchmarking GPT-5 on 400 real-world code reviews"

gsandahl — Fri, 08 Aug 2025 13:25:14 +0000

Most of the tasks have assessed with ground truth, occasionally helped with an LLM as a judge to assess the answer if the answer is a sentence and not an exact result.

Example: Given a long travel journal How many cities does the author mention? GPT-5: 12 Expected: 17

New comment by gsandahl in "Benchmarking GPT-5 on 400 real-world code reviews"

gsandahl — Fri, 08 Aug 2025 13:22:51 +0000

We are running task specific benchmarks across a number of categories (agentic tasks, context tasks, normalization tasks etc), and on our benchmarks we see Gpt-5 rating slightly below o3. But at a much lower cost.

See https://opper.ai/models

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 17:17:39 +0000

Please do and give us some feedback!

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 17:17:17 +0000

I think just how far you can go with examples has been an interesting learning! As these models have become smarter, they are also getting better at reasoning from examples and understanding intent. We will be publishing some research in the next few days!

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 13:43:11 +0000

No up to date demo video unfortunately :(

Sounds like a great use case though!

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 13:42:06 +0000

We have been thinking a bit about this, and one option would be to have some form of locally hosted runner. You can optimize the task in the cloud and deploy it locally. Something like that. It is possible to plug in custom models so technically feasible.

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 13:34:34 +0000

Yes that's possible! You can populate examples of great outputs to task specific datasets and have those be automatically populated to the prompt. More info here: https://docs.opper.ai/capabilities/learning

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 13:32:21 +0000

Thanks for the shout out!

New comment by gsandahl in "Show HN: Opper AI – Task-Completion API for LLMs"

gsandahl — Tue, 15 Jul 2025 13:08:53 +0000

Co-Founder here thanks for taking a look at Opper! I’m hanging around the thread all day, so feel free to ask anything, share feedback, or tell us where you’d like the product to go next

Schema Based Prompting: Structured Inputs for Predictable Outputs

gsandahl — Wed, 05 Mar 2025 10:04:38 +0000

Article URL: https://opper.ai/blog/schema-based-prompting

Comments URL: https://news.ycombinator.com/item?id=43264856

Points: 4

# Comments: 0

New comment by gsandahl in "Opperator: A composable agent to automate tasks on the web"

gsandahl — Fri, 14 Feb 2025 07:15:12 +0000

Its on that trajectory at least :)