Hacker News: frozenseven

New comment by frozenseven in "System Card: Claude Mythos Preview [pdf]"

frozenseven — Tue, 07 Apr 2026 19:42:45 +0000

Couldn't agree more. The "safest" AI company is actually the biggest liability. I hope other companies make a move soon.

autoresearch-genealogy

frozenseven — Thu, 19 Mar 2026 10:49:32 +0000

Article URL: https://github.com/mattprusak/autoresearch-genealogy

Comments URL: https://news.ycombinator.com/item?id=47437315

Points: 1

# Comments: 0

New comment by frozenseven in "Hive: A swarm of AI agents evolving code together"

frozenseven — Thu, 19 Mar 2026 07:07:10 +0000

GitHub: https://github.com/rllm-org/hive

Hive: A swarm of AI agents evolving code together

frozenseven — Thu, 19 Mar 2026 07:07:10 +0000

Article URL: https://hive.rllm-project.com/

Comments URL: https://news.ycombinator.com/item?id=47435892

Points: 1

# Comments: 1

New comment by frozenseven in "RetroAgent: From Solving to Evolving via Retrospective Dual Intrinsic Feedback"

frozenseven — Tue, 17 Mar 2026 04:28:21 +0000

Paper - https://arxiv.org/abs/2603.08561

RetroAgent: From Solving to Evolving via Retrospective Dual Intrinsic Feedback

frozenseven — Tue, 17 Mar 2026 04:28:21 +0000

Article URL: https://github.com/zhangxy-2019/RetroAgent

Comments URL: https://news.ycombinator.com/item?id=47408623

Points: 1

# Comments: 1

AutoResearchClaw

frozenseven — Sun, 15 Mar 2026 09:48:13 +0000

Article URL: https://github.com/aiming-lab/AutoResearchClaw

Comments URL: https://news.ycombinator.com/item?id=47385787

Points: 2

# Comments: 0

autoresearch-rl

frozenseven — Sun, 15 Mar 2026 01:36:02 +0000

Article URL: https://github.com/vivekvkashyap/autoresearch-rl

Comments URL: https://news.ycombinator.com/item?id=47383346

Points: 3

# Comments: 0

AutoContext: closed-loop system for improving agent behavior over repeated runs

frozenseven — Sat, 14 Mar 2026 00:56:15 +0000

Article URL: https://github.com/greyhaven-ai/autocontext

Comments URL: https://news.ycombinator.com/item?id=47372119

Points: 2

# Comments: 0

autoRL

frozenseven — Wed, 11 Mar 2026 22:12:48 +0000

Article URL: https://github.com/harshbhatt7585/autoRL

Comments URL: https://news.ycombinator.com/item?id=47342884

Points: 3

# Comments: 0

New comment by frozenseven in "I'm going to build my own OpenClaw, with blackjack and bun"

frozenseven — Wed, 11 Mar 2026 08:51:47 +0000

Cool project. Good luck!

New comment by frozenseven in "Ed Zitron loses his mind annotating an AI doomer macro memo"

frozenseven — Wed, 11 Mar 2026 08:01:27 +0000

>the stuff Zitron puts out assumes AI will continue to get better until the money runs out

This is categorically false. The big part of his "thing", from the first moment he ever started talking about AI, is the insistence that AI isn't getting better and very soon undergo a devastating collapse.

AutoKernel: Autoresearch for GPU Kernels

frozenseven — Wed, 11 Mar 2026 07:42:10 +0000

Article URL: https://github.com/RightNow-AI/autokernel

Comments URL: https://news.ycombinator.com/item?id=47332688

Points: 47

# Comments: 10

New comment by frozenseven in "Ed Zitron loses his mind annotating an AI doomer macro memo"

frozenseven — Wed, 25 Feb 2026 18:16:02 +0000

>The top open models are already good enough.

Rare opportunity for me to actually downplay frontier AI for a change. We can do a lot better. I think the next 6 months will be a stream of releases that shall leave all the current models in the dust. Opus 4.6 will be no more relevant than 3.5 Sonnet.

If this is the case, all bubble talk will have to be re-evaluated.