ThoughtAI
@Thought_THT
Followers
4K
Following
1K
Media
218
Statuses
956
ThoughtAI Adaptive AI Using Self-Aware Intelligent Software Agents. Telegram: https://t.co/jEETGqKlwu Discord: https://t.co/xl6qUlKYKy https://t.co/EYWW2ImihR
United States
Joined January 2018
🧵THOUGHT AI THREAD CONTEST!🧵 With less than 3 weeks till ARC-AGI2 submission, we wanted to have some fun! For your chance to win $THT simply write a thread on Thought AI. How to enter: - Follow @Thought_THT - Tag us in your thread - Reply here with the thread link Top 3
30
32
141
The models already struggling with arc-agi2 will struggle even more, if at all, with arc-agi3.
An interesting property of ARC 3 is that it is more accessible to children than ARC 1 & 2, while being much more difficult for current AI systems
3
7
52
On it 🫡
Now it's up to us to refine and scale symbolic AGI to save the world economy before the genAI bubble pops. Tick tock
3
7
55
One more nuclear power plant, bro. Please bro.
Power demand from AI data centers to quadruple in 10 years, per Bloomberg:
6
4
40
Sam Altman’s OpenAI will use as much power as New York City and San Diego combined, per FORTUNE
555
1K
13K
😳
GPT-5 consumes like a country, it ranks #81 in global electricity use GPT-5’s estimated annual electricity consumption is about 16.4 TWh; it consumes way more energy than Slovenia as a country.
4
6
30
🕵️♂️ Unearthed this gem from our archives! Watch our CEO Prof. Andrew Hacker dive into Agent-based Models (ABM) – history, trends, nature-inspired insights, emergent properties, & why ABM is the path to super-intelligent AI & digital life. 🤖🌍 👉
6
19
65
Would we be seeing headlines like this regularly if any of the big labs was close to AGI? Of course not.
32
58
455
Has the LLM high-water mark arrived? Probably a good time for us to soon compete for ARC-AGI2 benchmark with a completely novel approach…
We were promised AGI, ASI, personal superintelligence. Instead, we will get infinite slop machines that turn us into dopamine-addicted zombies.
3
21
57
We suspect that, seemingly out of nowhere, a novel, patented approach will surprise just about everyone in a month or so. @Thought_THT will serve up what they've been cooking for years. Only recently has their approach to the problem been given any attention.
Grok 4 Heavy is a version of Grok 4 that uses multiple agents. Instead of relying on just one model, agents work together simultaneously on the same task. After they produce their results, they compare their findings and agree on a final answer. In Humanity's Last Exam, Grok 4
5
26
54
How do you go against the Arc-AGI benchmarks? As the Arc-Agi2 leaderboard heats up and we near its conclusion… have a go yourself and see if you can solve the daily puzzle! Post your efforts into our Discord and TG if you manage to solve and let us know how you did. There
4
12
44
In 2023 there was a big uptick in companies marketing their products as “AI powered”. 2026 will be the year of companies marketing their products as “AI free” (a trend already underway)
82
198
1K
OpenAI and all major AI companies betting on gigawatts of AI compute.. It'll be ironical if a new kid on the block makes all of it redundant by figuring out a general intelligence that runs on 20 watts of power (like human brain does).
443
171
3K
🕵️♂️ Unearthed this gem from our archives! Watch our CEO Prof. Andrew Hacker dive into Agent-based Models (ABM) – history, trends, nature-inspired insights, emergent properties, & why ABM is the path to super-intelligent AI & digital life. 🤖🌍 👉
6
19
65
It took over a decade for Jensen’s thesis to play out. It’s been over a decade since we filed our patents for Agentic AI and our thesis is about to play out.
In 2009, Jensen explained CUDA and estimated its market size, $NVDA stock price was at ~$0.2 split adjusted. $10k investment in NVIDIA would turn into $8.8M The GOAT.
9
25
63
@Thought_THT has started a public challenge to build more advanced AI (the ARC AGI Race). If you want to learn how to buy $THT on Coinmetro, here’s a short, step-by-step video. 🎥
0
10
36
Time for a real Thought process. https://t.co/MlkUw1li88
arstechnica.com
Chain-of-thought AI “degrades significantly” when asked to generalize beyond training.
1
11
50
MARK ZUCKERBERG: “We are going to spend aggressively. Even if we lose a couple hundred billion, it would suck, but it’s better than being behind the race for super intelligence.” To even hear Zucks say this…it shows how important they believe the opportunity is. Cap-ex is
385
669
9K
💡Did you know? Thought AI holds a powerful portfolio of patents that make its core technology nearly impossible to replicate. 🔒 We’re embedding intelligence right into data, powering real-time decisions across distributed systems. These innovations put us in a league of our
3
30
89