Y Combinator
Region: US
Friday 30 October 2026 20:01:19 GMT
74251
3604
77
482
Music
Download
Comments
marianelarojas64 :
Nothing beats Claude still, I wish there was something open source
2026-10-31 06:56:09
4
unprotectedsets :
I applied to you guys once and you turned me down :(
2026-10-31 00:29:26
38
R :
AI is going to be like S3, there is no moat. the level of intelligence is sufficient right now, most companies don’t need to solve millennium problems. the moat is in the harness.
2026-10-30 23:11:59
40
Barza :
i’m working on this
2026-11-01 21:51:16
1
Pooty Patoo :
The open weight models are benchmaxxed. Distilled from the best models, they do well on benchmarks but poorly in real world work.
2026-10-31 00:09:54
4
Dubby :
love the way you broke this down
2026-11-01 14:46:39
0
alex0000 :
How can you monetize local models? I mean for startups that try to sell such solutions (local AI). Everything running on the customer’s hardware can be reverse engineered and pirated
2026-10-31 12:25:38
0
Daniel Marcus :
I’m building the data LLM’s depend on to give great answers. Who wants to fund the shovel during the Gold Rush.
2026-11-01 09:59:18
3
S.E Global |Systems & Strategy :
This is where the next competitive advantage in AI will emerge: not just access to intelligence, but ownership of the systems that make intelligence improve over time... A model alone is not institutional intelligence. The real value comes from capturing domain expertise, structuring proprietary data, evaluating outputs against real business workflows, and continuously learning from operational feedback. The companies that master this will move beyond simply consuming AI to building intelligence around how their businesses actually operate.. The future is not just about who has the most powerful model, but who has the most valuable institutional knowledge embedded in their systems...😏
2026-11-01 01:53:45
2
Sniper 217 :
Why he never even blink for almost a minute
2026-10-31 02:27:43
0
P Tid :
yes yes yes yes. currently running LoRa boxed to Qwen 4! Fable handles the fine tuning after ingesting Terry's of data y'all and everyday this tiny thing running on a laptop with a 5060 sounds more and more like my customer service reps everyday
2026-10-31 08:57:55
0
👤Dr TkHarris Mental Wealth ❄️ :
Important for us in medicine and therapy. Nobody is helping me make a standalone safe AI for my therapeutic work
2026-11-01 13:35:42
1
Poetaytoe :
that means maintenance, why not an AI that learns from that data while the company still owns its own data an not the AI company. 😅 plug n play an it learns.
2026-10-31 01:54:29
0
Tiktoktik :
Rent space or rent models it’s the same
2026-10-31 00:17:46
0
EIAI :
Sylvanity (https://sylvanity.ai) does that with our own models.
2026-10-30 21:51:44
3
BS :
Owning intelligence is the "idea" I've been working on for a while. Happy to get in contact!
2026-10-30 23:11:18
0
Derric :
be careful if Dario sees you using ram he will not be happy
2026-10-31 12:02:17
0
AIguy :
Well explained this how you explain what AI software developers do
2026-10-31 11:31:55
1
codeeveryday :
devs are now dependent on AI subscriptions
2026-10-31 02:06:00
1
ramdane :
well I want to be one of the people who does that
2026-10-30 21:57:30
1
Field Notes & Future Systems :
People continue to miss one key point. If you are relying on external AI you are relying on a “workforce” who is fundamentally conflicted to help someone else’s profits, not yours. It will help just enough while goal seeking to maximise someone else’s profits. So 100% it will be not just your own AI but it will be on-prem. China is also making no secret of that fact it will provide open source and for free what others are charging huge amounts for.
2026-10-31 09:59:45
0
Michael :
https://openshard.dev is doing this !
2026-11-01 00:15:46
0
Mark :
I thought this for a while but the frontier models are just pulling ahead in benchmarks and real work
2026-10-31 16:32:59
0
user3855845335797 :
What if I’m looking into doing just that?
2026-10-30 21:07:31
0
x ss4 :
it's the infrastructure that hinder the cost efficiency, you expect some startups will beat nvidia, intel, etc in the economic scale?
2026-11-01 03:19:37
0
To see more videos from user @y_combinator, please go to the Tikwm
homepage.
