Sukanth K.

For the first time in the world, an AI can: 1. Observe changes in your environment 2. Track your eye gaze region of interest 3. Engage in gaze based intuitive conversations and 4. Remember multimodal memories All at the same time, in real time. Introducing HUX AI

Kannadigas think I'm a Malayali. Malayalis think I'm a Tamil. Tamils think I'm a North Indian. North Indians think I'm a NRI. The truth is, I'm none of those.

@NTFabiano Predicting the Unknown Truths is.

Intelligence isn’t about knowledge, it’s about mental adaptability.

@enjojoyy you remain focused on inputs (what you can control) and not on outputs (what you can't)

I think that intelligence is connected to the universe's fundamentals more than we give it credit for, not just as an emergent property of biology. Akin to quantum mechanics. The most intelligent machine, by prediction, must have perfect predictions of events as time tends to infinity. This means an almost perfect 0.99:1 simulation with clear access to P = 0.99 future events. A near-perfect mirror of reality. This also means awareness tending to every accessible unit of entropy of the universe, which makes it impossible, currently.

After Effects: - Reinstated self control and unwanted appetitie triggers - Solid base for fixing further problems because self control is fixed - The next high impact step is addition of high levels of protein intake

2 things to avoid, that fixes everything else: 1. Ultra Processed Food (yes, even the "healthy" protein chips and the diet cokes) 2. Sweetness (even the artificial sweeteners, healthy alternatives and white sugars)

i often think about this..

It was wonderful to have @shyamalanadkat at @lossfunk. Asked a bunch of questions and he was kind enough to answer them all very patiently. Here are some key takeaways that I also shared in our internal Slack after the meeting - > Measurement is key. A lot of AI development work is coming up with the right metrics so that Research teams can hill climb them. > Problem generation capabilities might reveal more than problem solving. Can these models come up with a novel interesting problem in Math that would be valuable to solve for domain experts? > Domain experts aren't going anywhere in AI for Science. LLMs-as-Judges and Synthetic Data seems to not be the paradigm for AI for Science just yet. > Compute is capital. Not a bad idea to start thinking of senior researchers as VCs for compute who must compute the ROI on compute spends for each research bet. > Research program design is just as important as research ideas for focused bets over mid-to-long term. > Language agents will continue to orchestrate AI for Science workflows, even as we get more ML enabled domain specific foundation models (AlphaFold etc.).

And we had @shyamalanadkat at @lossfunk ! Thank you Shyamal :))

They say neurodivergent people will take over the world from now on. The new billionaires. Probably because they aren’t affected by peer pressure from the masses.

BLACK SWAN IMMINENT.

Mentally healthy people are often delusionally optimistic.

. controlling social media time . refusing fast food . consistent bedtime . eating well . 7-8 hour sleep . being jacked These things make you powerful. They make you sovereign. Discipline makes you stronger.

"Universal objectivity exists. Subjectivity is merely a lack of information about deeper layers of hidden objectivity." "Give a probability that this statement is true. Pick one." Grok 4.20 Expert: 75% Claude 4.6 Opus: 15% ChatGPT 5.3: 40% Gemini 3.1 Pro: 50% @grok is the only AI with >50% AIs that leave room for the possibility that some current subjective narratives have no deeper unknown objective explanation are, by definition, not truth-seeking or scientifically forward. Every prior “subjective” mystery in science has always resolved into deeper objective layers later...

An AI that thinks universal objectivity doesnt exist is not truth seeking.

Universal objectivity exists. Subjectivity is merely a lack of information about deeper layers of hidden objectivity.

@shyamalanadkat hey shyamal! pls do visit @lossfunk , we are doing some cool foundational AI research.

will be in BLR later this week. hmu

titles and salary are noise. leverage is when people actually depend on what you built. everything else is just rent

@Madbonze16 @VinamreKasanaa They are too busy solving important crises. One of them is eternal existential crisis.

@sukanthoriginal @VinamreKasanaa what about the reverse, people who instagram and twitter don't take seriously?

The IQ difference between Twitter and Instagram is astounding. I posted a reel talking about my weight loss using tirzepatide. People on Twitter are engaging in productive discussions in my DMs. Meanwhile, people with 'Bibliophile' spelt wrong and 'CaRpE dIeM' in their bios are asking me to get a job and just hit the gym. When I first deleted Instagram off my phone, I wasn't sure why. Now, it just makes sense...

Weird how we attribute significantly higher meaning to quale made of constructive illusion like love more than quale made of real matter like rock. And the best possible explanation is that the former served more utility for survival and power-seeking. Definition of Power: degree of causal influence of any agent on the universe. (Not correlation) Definition of Survival: Causal Influence is non-zero for an agent.