@DeryaTR_ A model randomly matching patterns to come up with a "novel" hypothesis is hardly surprising is it? Let's not conflate an AI model performing pattern predictions based on data with the notion that there is not a plateau in actual progress. That is the most basic function of what

@lucasbaker You have no idea. But anyways, you shall see soon ;)

@DeryaTR_ Gary would actually be a wonderful contributor to the scene if he stopped being so whiny about deep learning. His concept/symbol networks probably do have a lot of value as a side/sub-component to current AI architecture.

@rohanvisme This isn’t his concept per se, but back in the day he did make some conceptual contributions. That was when we were still friends; then he moved to the dark side (or, from his perspective, we did). 😅

@DeryaTR_ That's not "intelligence". It is statistical pattern matching and brainstorming.. those ideas are analyzed and filtered by professional scientists for sanity. No one is saying LLMs aren't useful... but they aren't AGI.

@Jay_Redman_68 Do you realize that I am actually scientist working on cancer?

@DeryaTR_ How deep are you going to bury Gary? 😂

@markopolojarvi That’s a great pun 😂 No limit on the reverse side either! 😅

@DeryaTR_ Remind me again why we continue to listen to Mr. Marcus.

@Fedesco5 To create new memes for the worst predictions in the age of AI 😅

@DeryaTR_ it’s preclinical. you used to be a scientist. and you used to know what meant. how many candidates like this are there? how many survived clinical trials?

@DeryaTR_ The saddest part about his misplaced delusional glee is that he seems to be relishing the idea that AI drug discovery and the millions of lives it could save or make healthy won’t happen.

@DeryaTR_ He has said he has won every year LMAO 😂🤣

@DeryaTR_ I learned that people would rather debate the nuances of “novel discovery” than be excited about curing cancer.

@DeryaTR_ "in living cells", meaning in vitro. Remind me again the success rate of drugs from in vitro, to in-vivo, to clinical trials?

@DeryaTR_ AI currently at level 4 (innovator stage), achieves many breakthroughts recently in various scientific fields. And we are still early..

@DeryaTR_ Gary Marcus talking about how AI is about to hit a wall every year.

@DeryaTR_ "AGI" vs "AGFI" "AGFI" -> Artificial General Fine-Tuned intelligence i.e. benchmarks are gamed; models are fine-tuned to ace them. That's not intelligence. True AGI: AI self-improves linearly every week in all domains; sublinear growth is also acceptable initially.

@DeryaTR_ GPT-5 Pro: Imagine long tasks like walking a tightrope for a thousand steps. Two things prevent you from falling: you place each step carefully, and when you start to wobble, you don’t carry the wobble into the next step; instead, you regain your balance. https://arxiv.org/html/2509.09...

@DeryaTR_ @grok this means you are doomed too!

@DeryaTR_ Still waiting for him to come up with a paradigm changing neuro symbolic architecture that can match what is currently out there Until then, I've tuned everything else he says out

@DeryaTR_ He will NEVER admit it, if anything he will double down despite all evidence and advancements showing contrary to his statement.

@DeryaTR_ I guess humanity win more if we get more humble and learn how to appreciate our own mistakes. And stay open to new mistakes.

@DeryaTR_ Clearly, what started as hope is about to turn into chaos for them.

@DeryaTR_ Good morning 🤝 gave you a follow

@DeryaTR_ Good morning 🌄☕

@DeryaTR_ In my humble and untrained opinion, we’re not close to AGI, but we are creating powerful narrow AIs that will solve cancer, aging, and other problems.

@DeryaTR_ Get up, Derya. The group chat needs you. 💬📱 How is your Thursday going?

@DeryaTR_ Morning! Remember that even the smallest step is still a step forward. 🧡 Have a heartwarming morning.