Vijay

“यूँही आबाद रहेगी दुनिया,
हम न होंगे, कोई हम सा होगा”

Follow @signaloverx

[Vidit Gujrathi @viditchess]

Chess engines tell you the best move.

But grandmasters are human, they don’t always play it.

So I built "Kibitz": a human move predictor for chess broadcasts. I trained this model on my Nvidia RTX 5080.

Then I made it run as a business by itself.

A channel buys the overlay, Hermes onboards them, charges via @stripe test mode, runs the broadcast, narrates with @NVIDIAAI Nemotron, tracks inference cost, and books its own P&L.

I build. Hermes operates.

This is my demo and entry for the @NousResearch × @NVIDIAAI × @stripe Hermes Agent Accelerated Business Hackathon.

[The Other Guy @Gyan49485]

A world championship is a responsibility too. You can't just enjoy the fruits of it without carrying it's burden.

The responsibility of being WC also includes the responsibility to lose it to the next person, whenever it comes. Not doing so is disrespecting the title itself.

[Vijay @BazingaBanter]

ABE CHUTIYE GANDU LOG

[News Algebra @NewsAlgebraIND]

French and Russian revolutions dropped from Class 9 books by NCERT.

[Vijay @BazingaBanter]

@utkarsh_rawat0 @KanojiaPJ @IDF @netanyahu Asking a foreign politician to bomb India? Are you a Yogi supporter?

[Vijay @BazingaBanter]

@SwiggyCares 240252837411134

[Vijay @BazingaBanter]

Hi @SwiggyCares my order got delayed by a lot. Can you please check

[Signal Over Noise @signaloverx]

One of the shortest serious mathematics papers ever published destroyed a 200-year-old conjecture.

Not with a long proof.
Not with a deep new theory.

Just a single counterexample found by a computer search on the CDC 6600; one of the most powerful computers of its time.

That's one of the beautiful things about mathematics: if a statement claims something is always true, one exception is enough.

A centuries-old idea can fall to a few carefully chosen numbers.

Paper: https://t.co/zIuaQm3NJi

[Signal Over Noise @signaloverx]

1/ One of the most important decisions in human history was made by a man most people have never heard of.

His name was Vasili Arkhipov.

In 1962, his decision may have prevented a nuclear war.

🧵

[Vijay @BazingaBanter]

Happened today...

I was riding my scooty on a narrow lane when a guy suddenly appeared from a house gate straight onto the road

I somehow avoided a collision

Rhetorically, I asked my brother..

"Whose mistake was it?"

He replied

"Nehru"

😂😂😂😂😂

[Aunindyo Chakravarty @Aunindyo2023]

Never, ever, claimed to be a journalist. I was always just a mere news professional. Journalists are much higher than me. You, however, are a dalaal.

[Vijay @BazingaBanter]

Please no

[Chennai Updates @UpdatesChennai]

🚨Tamil Nadu to consider bidding for World Chess Championship. Final decision to be taken post consultation with CM Vijay... #TN #Chess ♟️

[Signal Over Noise @signaloverx]

Consistency is powerful because humans mistake repetition for truth.

That’s why propaganda works.

Repeat a claim long enough:

elections are rigged
reservation is anti-merit
minorities are the problem
welfare schemes are all bad

…and eventually familiarity starts feeling like evidence.

Psychologists call this the “illusory truth effect”:
repeated statements become easier to process, and the brain often mistakes that ease for accuracy.
Modern misinformation ecosystems, especially algorithmically amplified political media exploit this constantly.

The goal is rarely to prove something. It’s to repeat it until skepticism becomes exhausting.

Study:
“The Illusory Truth Effect” Fazio et al., 2015
https://t.co/XkjNbF8KAZ

[chess24 @chess24com]

Praggnanandhaa wins a crazy game to end Sindarov's 53-game unbeaten streak in classical chess! https://t.co/aMbocrEpRC

#GrandChessTour

[Signal Over Noise @signaloverx]

Accurate!

[Vijay @BazingaBanter]

@shashank_ssj @HMOIndia @DelhiPolice Abe chu

[Vijay @BazingaBanter]

@barandbench Disgusting

[Signal Over Noise @signaloverx]

People think GPT overuses em dashes (—) because “AI writes weird.”

But the real reason says something disturbing about the internet itself.

GPT was trained on millions of articles, blogs, essays, newsletters, thinkpieces, and “smart sounding” online posts written by humans.

And in those ecosystems, writing style directly affects money and status:
• more clicks
• more retention
• more authority
• editor approval
• virality
• intellectual branding

Over time, professional internet writers evolved toward a very specific rhythm of writing:

“not this — but this”

Why?

Because em dashes do something psychologically powerful:
they mimic human thinking in real time.

They create pauses.
They compress ideas smoothly.
They make sentences feel nuanced and conversational at the same time.

Example:

“Humans don’t fear AI — they fear irrelevance.”

That structure became massively overrepresented in polished online writing.

So GPT statistically learned:
“this punctuation pattern correlates with writing humans rate as intelligent.”

Then came RLHF (human feedback training), where reviewers consistently rewarded answers that sounded:
• thoughtful
• nuanced
• fluid
• natural

So no, nobody programmed GPT to love em dashes.

It inherited the rhetoric of professional internet humans trying to sound smart.

The funniest part?

humans created a prestige writing style → AI mastered it → humans abandoned it to escape association with AI

[Signal Over Noise @signaloverx]

To transport natural gas across oceans, it’s cooled to -162°C until it becomes liquid.

Its volume shrinks ~600 times

Now the interesting part:

Some of the LNG slowly evaporates during the voyage and many tankers reuse that gas to help power the ship itself

Sounds fictional?!

[Signal Over Noise @signaloverx]

Beautiful

[Vijay @BazingaBanter]

@amit_ganguli These people are digital terrorists