alansaber

9 ms·

241.
by alansaber 2mo ago
no thanks. i use AI.
242.
by alansaber 2mo ago
Code mode getting some love.
243.
by alansaber 2mo ago
Agreed that it is a killer feature. US models obfuscate the COT (to A. make it look better and B. combat distillation) but > and the raw trace is fairly hard to reason about > but I still think this kind of feature is a big step in th
244.
by alansaber 2mo ago
Growth hacking
245.
by alansaber 2mo ago
Power users will always want custom tools. But most users aren't power users
246.
by alansaber 2mo ago
You don't think we already do? What's missing? Better inter-agent communications? Better unit tests and linting?
247.
by alansaber 2mo ago
Even before AI this was the best way to sustain attention/effort imo
248.
by alansaber 2mo ago
This has always been true. No?
249.
by alansaber 2mo ago
Not sure I see what would displace humans from the top. AI keeps getting better at the middle bit. Sure, that means you can be lazier with instructions from the top. But surely we see diminishing returns from that- you can only RL so hard,
250.
by alansaber 2mo ago
Probably because with fewer tabs, you actually get to do the deeper thinking that you enjoy? With more, you're constrained to quickly converging on the best way to keep the plate spinning
251.
by alansaber 2mo ago
"Fewer iterations for materials science discovery" is a good spin. Closing the computational>experimental loop is the main challenge. This is the focus of my past research group, there is definitely potential, best of luck!! I
252.
by alansaber 3mo ago
Isn't quite a lot of modern smartphone photography post-processing? Yes camera SAAS is extremely funny but I can kinda see it maybe
253.
by alansaber 3mo ago
Valuations are significantly leveraged compared to actual user metrics. My impression of Manus was that, like Perplexity, some users were using it to get around firewalls/regulation.
254.
by alansaber 3mo ago
Neat.
255.
Multi-agent version control with reactive feedback(lexifina.com)
1 points by alansaber 3mo ago|0 comments
256.
by alansaber 3mo ago
The article lost me when it implied that verbose drivel is actually intrinsically superior rather than a way to hedge bets
257.
by alansaber 3mo ago
Not sure what happened in the blog, but I quite enjoyed the mindmap in the right panel
258.
by alansaber 3mo ago
There are many ways for a business to survive outside of innovation
259.
by alansaber 3mo ago
Cool, we do the same thing, but we also denote when a line is "AI generated but was modified by a human" (aka human made anything upwards of a 1 character change).
260.
by alansaber 3mo ago
Trained on short form gratification loops? Just the opposite. They're the path of least resistance. It was thermodynamically inevitable that short form content would dominate.
261.
by alansaber 3mo ago
Fake it til you make it
262.
by alansaber 3mo ago
Given the amount of raw compute going into models it would be more surprising if we couldn't get events like this
263.
by alansaber 3mo ago
They'll set up guardrails but I believe the point is better code uae / better long running tasks > inevitable that cyberattacks will be easier
264.
Solving Legal Problems with Dags(lexifina.com)
4 points by alansaber 3mo ago|0 comments
265.
by alansaber 3mo ago
Live by the scraper, die by the scraper
266.
by alansaber 3mo ago
I get the market need for AI detection, but the fundamentals of it are so woo-woo that I despair
267.
by alansaber 3mo ago
Fair enough.
268.
See inside your agents to fix them(lexifina.com)
2 points by alansaber 3mo ago|0 comments
269.
by alansaber 3mo ago
The ragebait > AI pipeline remains strong
270.
by alansaber 3mo ago
Concern for celestial events is a bit like one of your gut bacterium being concerned with your work schedule

More ›