Saoud Khalifah
@SaoudKhalifah
Followers
3K
Following
2K
Media
91
Statuses
1K
Building the future | ex-Founder and CEO, Fakespot (acq'd)
Matrix
Joined January 2021
Plato & Aristotle may have more to teach LLMs than we think. Dialectic ≈ Algorithms. Both seek truth—step by step. What if the future of AI lies in the past?. New post →
14
2
63
I do wonder if the tariffs on imports (they are reduced according to today's report) are going to reduce consumer culture or do the opposite.
0
0
0
Seems like what could have been a singular startup yesterday, should be an AI agent tomorrow. One that browses, plans, books, updates, messages, and closes transactions on its own. It feels inevitable, right?. However, there are problems. Agents amplify mistakes. One wrong
1
0
3
When i see stuff like this it always reminds me of, correlation != causation.
0
0
5
In an age of spellcheck and LLMs, how did @Zoom miss this typo in an email *about* an AI companion?
3
0
8
The future of AI is to make the non-deterministic (lack of context) -> more deterministic with (increased level of detail/context).
1
0
9
Current AI is quite flawed, but I think this show-cases the potential for how super personalized attention/care in a 1-on-1 chat with an AI bot is a massive differentiator versus the human equivalent. You may share things at anytime with no need to create any appointments or rush
0
0
6
Was trying to join in with the memes today, 2nd image is completely off but absolutely hilarious.
1
0
5
100% on the money here.
if you're a vibe coder rolling your own auth, there's no way in hell i'm signing up for your platform.
3
0
8
I can confidently say that I've never seen folks use em-dashes prior to ChatGPT and Claude.
1
0
13
I wonder how many of them were coding/CS books.
Anthropic purchased millions of physical print books to digitally scan them for Claude
0
0
6
Humans are building AI, and AI is building humans. The social implications of this are still not fully comprehended.
3
0
16
The dichotomy that I have observed with AI models is between deterministic algorithms and non-deterministic AI models. Looks like the pendulum is swinging in the direction of the acceptable "randomization" - protected by algos.
0
0
8
The intricacies of the font design on stone back by the Romans in Ancient London is just amazing (zoom in to see how beautiful it is).
1
0
5
Smart.
0
0
3
My bet is that @miramurati and @ilyasut are working on next generation of foundational models. let's wait and see.
0
0
3
Damn, I must admit this is pretty dope.
The last time intelligence exploded on Earth, it wasn’t exactly amazing for everyone else. Here’s a fable about risks from transformative AI (made with Veo 3)
0
0
4
Interesting seeing @karpathy compare LLMs to utilities. In a way this is a self-referential loop, since each answer "drawn" from the grid has the potential to rewire the grid by updating weights. Humans are the feedback signal amplifying the voltage and also part of the grid.
0
0
8