sshh12

sshh12 on Hacker News

Classifying Emails with Jev and the OpenAI Decisions API

Posted by sshh12 on Thursday October 08, 2026 @11:30AM
from the builders.abnormal.ai dept.

https://builders.abnormal.ai/p/classifying-emails-with-jev-and-the

( Read More... | no comments | on HN )

Re:Every SaaS business will become a harness around a model on Friday October 02, 2026 @07:16PM (#49939757)

Yeah 100% agree

Re:Every SaaS business will become a harness around a model on Friday October 02, 2026 @06:07PM (#49939209)

yes very unserious software: https://abnormal.ai/

Nanojev: A minimal Jev-style decision model in 200 loc

Posted by sshh12 on Tuesday September 22, 2026 @11:59AM
from the github.com dept.

https://github.com/sshh12/nanojev

( Read More... | no comments | on HN )

I Asked 100 Agents to Hack Me

Posted by sshh12 on Tuesday September 08, 2026 @10:23AM
from the blog.sshh.io dept.

https://blog.sshh.io/p/i-asked-100-agents-to-hack-me

( Read More... | no comments | on HN )

The Harness Is the Company

Posted by sshh12 on Monday August 24, 2026 @10:22AM
from the blog.sshh.io dept.

https://blog.sshh.io/p/the-harness-is-the-company

( Read More... | 1 comment | 2 points | on HN )

Mitigating Reward Hacking as Institutional Design

Posted by sshh12 on Thursday August 20, 2026 @09:22PM
from the beren.io dept.

https://www.beren.io/2026-08-17-Mitigating-Reward-Hacking-as-Institutional-Design/

( Read More... | no comments | 2 points | on HN )

Show HN: Launder, a puzzle you can only win by removing SynthID watermarks

Posted by sshh12 on Wednesday August 19, 2026 @10:31AM
from the show-hn dept.

https://launder.sshh.io/

( Read More... | no comments | on HN )

Using AI in 2026 (Coding, Writing, Learning, Assistant-Ing)

Posted by sshh12 on Monday August 17, 2026 @10:11AM
from the blog.sshh.io dept.

https://blog.sshh.io/p/how-i-use-ai-in-2026-coding-writing

( Read More... | no comments | 2 points | on HN )

Re:Exploring Claude/GPT Knowledge Cutoffs and Pre-Training Timelines on Monday August 10, 2026 @04:28PM (#49249254)

I would have thought that as well but at least for coding vs world event facts I didn't see obvious differences in cutoffs

Exploring Claude/GPT Knowledge Cutoffs and Pre-Training Timelines

Posted by sshh12 on Monday August 10, 2026 @10:20AM
from the blog.sshh.io dept.

https://blog.sshh.io/p/exploring-claudegpt-knowledge-cutoffs

( Read More... | 24 comments | 4 points | on HN )

How Can LLM RL Work Despite Information-Theoretic Inefficiency

Posted by sshh12 on Saturday August 08, 2026 @06:32PM
from the beren.io dept.

https://www.beren.io/2026-07-26-How-Can-LLM-RL-Work-Despite-Information-Theoretic-Inefficiency/

( Read More... | no comments | on HN )

Make Every PR Prove Itself

Posted by sshh12 on Friday July 31, 2026 @02:00PM
from the builders.abnormal.ai dept.

https://builders.abnormal.ai/p/make-every-pr-prove-itself

( Read More... | 1 comment | 6 points | on HN )

Show HN: A QR code that plots where you scanned it from

Posted by sshh12 on Sunday July 26, 2026 @02:56PM
from the show-hn dept.

https://3dqr.sshh.io/

( Read More... | no comments | on HN )

Designing Software for Software Factories

Posted by sshh12 on Saturday June 13, 2026 @06:34PM
from the blog.sshh.io dept.

https://blog.sshh.io/p/designing-software-for-software-factories

( Read More... | no comments | 2 points | on HN )

Waypoint-1.5

Posted by sshh12 on Saturday June 06, 2026 @06:48PM
from the over.world dept.

https://over.world/blog/waypoint-1-5

( Read More... | no comments | 3 points | on HN )

Re:Elevated error rates on requests to multiple models on Saturday May 16, 2026 @02:23PM (#48162515)

Had codex whip up a ./claude-via-codex.sh for resuming claude cli sessions in the meantime:

https://gist.github.com/sshh12/f2c402db02eb00982100f645519bd...

WikiProject Editor Retention

Posted by sshh12 on Friday May 15, 2026 @09:16PM
from the en.wikipedia.org dept.

https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Editor_Retention

( Read More... | 1 comment | on HN )

Re:AI Productivity Fails on Sunday May 10, 2026 @09:44PM (#48090149)

Oof well I guess I'm personally a fan of buzz feed listicles. Thats been my style pre-AI.

Re:AI Productivity Fails on Sunday May 10, 2026 @09:39PM (#48090111)

Any issues with the content itself?

It shouldn't be that surprising that a post advocating for using more AI is written using AI.

AI Productivity Fails

Posted by sshh12 on Sunday May 10, 2026 @07:43PM
from the blog.sshh.io dept.

https://blog.sshh.io/p/how-ai-productivity-fails

( Read More... | 7 comments | 3 points | on HN )

AI drives companies to reorganize around loops, not functions

Posted by sshh12 on Sunday March 29, 2026 @04:14PM
from the blog.sshh.io dept.

https://blog.sshh.io/p/the-transposed-organization

( Read More... | no comments | on HN )

Show HN: Brw – Browser automation for Claude Code agent teams

Posted by sshh12 on Saturday March 07, 2026 @04:56PM
from the show-hn dept.

https://github.com/sshh12/claude-plugins/blob/main/plugins/brw/README.md

( Read More... | no comments | on HN )

Plan.md

Posted by sshh12 on Saturday March 07, 2026 @11:17AM
from the github.com dept.

https://github.com/abnormal-ai/claude-plugins/blob/main/plugins/spec-tool/skills/build-spec-tool/references/plan_template.md

( Read More... | 1 comment | on HN )

Re:Plan.md on Saturday March 07, 2026 @11:17AM (#47288927)

Hi all, sharing a template we started using for AI-based spec development in case it's useful

Revealed Preferences

Posted by sshh12 on Monday March 02, 2026 @07:59PM
from the writing.nikunjk.com dept.

https://writing.nikunjk.com/p/revealed-preferences

( Read More... | 1 comment | on HN )

Our Design Docs Write Themselves

Posted by sshh12 on Monday March 02, 2026 @11:12AM
from the abnormalbuilders.substack.com dept.

https://abnormalbuilders.substack.com/p/our-design-docs-write-themselves

( Read More... | no comments | 2 points | on HN )

Re:Reverse-engineering how Claude's Chrome extension controls the browser on Saturday February 28, 2026 @07:54PM (#47202393)

slack: https://gist.github.com/sshh12/4cca8d6698be3c80e9232b68586b7...

Re:Reverse-engineering how Claude's Chrome extension controls the browser on Saturday February 28, 2026 @07:48PM (#47202352)

netflix: https://gist.github.com/sshh12/dda3a89514f850c459380b18b1f7e...

Reverse-engineering how Claude's Chrome extension controls the browser

Posted by sshh12 on Saturday February 28, 2026 @07:16PM
from the gist.github.com dept.

https://gist.github.com/sshh12/e352c053627ccbe1636781f73d6d715b

( Read More... | 3 comments | 2 points | on HN )