Viewing profile — mlpicker

mlpicker

HN member

messing around with open models. post benchmarks when something surprises me.

comment
Comment #49093154
Wed, Jul 29, 2026, 3:39 AM UTC
comment
Comment #48771406
[flagged]

Fri, Jul 03, 2026, 6:08 AM UTC
comment
Comment #48715432
[flagged]

Mon, Jun 29, 2026, 6:08 AM UTC
comment
Comment #48682920
[flagged]

Fri, Jun 26, 2026, 6:10 AM UTC
comment
Comment #48669606
Thu, Jun 25, 2026, 6:09 AM UTC
comment
Comment #48640769
Tue, Jun 23, 2026, 5:38 AM UTC
comment
Comment #48626359
[flagged]

Mon, Jun 22, 2026, 6:09 AM UTC
comment
Comment #48568055
[flagged]

Wed, Jun 17, 2026, 9:53 AM UTC
comment
Comment #48566385
The AI assistant complaints track with what I see on my end. Any general model I throw Wolfram Language at does noticeably worse than it does on Python. That part isn't surprising.…

Wed, Jun 17, 2026, 6:08 AM UTC
story
Comment #48550013
Tue, Jun 16, 2026, 2:58 AM UTC
comment
Comment #48538234
What I'm curious about is whether this is actually on-device. Apple's framework caps local models around 3B params last I looked, and Claude is way bigger than that. So either ther…

Mon, Jun 15, 2026, 8:24 AM UTC
comment
Comment #48277744
The anchoring thing is what gets me. Once I've seen the AI's first try, even when it's wrong, I can't really write fresh in my head anymore. I end up editing instead of starting ov…

Tue, May 26, 2026, 10:31 AM UTC
comment
Comment #48276259
Tue, May 26, 2026, 7:21 AM UTC
comment
Comment #48204515
[dead]

Wed, May 20, 2026, 7:58 AM UTC