Aditya Jha

Founder and CEO @ Head of Mid Frequency @ Tower Research, VP Quant Research @rbccm, Quant Trading @jpmorgan

[Aditya Jha @36Jha]

@nichochar I’m in the same boat - running a company and have a baby daughter - but the universe here seems to have been restricted to Americans. Look out to other parts of the world and you’ll find plenty of examples. It’s all in the value system.

[Aditya Jha @36Jha]

If you’re on this list, I have just the thing for you 😉

[Kalshi Finance @Kalshi_Finance]

JUST IN: Nearly 75% of active fund managers failed to beat the S&P 500 over the last 12 months, per WSJ

[Aditya Jha @36Jha]

Strategies run for under a dollar now!

[Itoflow @itoflowai]

In its RLM experiments Prime Agent has:
- found lower performance on math-python
- worse DeepDive results without task-specific guidance
- higher latency in every tested environment

The Continual Harness paper evaluates Pokémon environments, not much of which translates to governed financial workflows.

So our push on Quantitative Harness continues. The game we're solving for is the toughest of them all.

[Aditya Jha @36Jha]

Luna may help us cut down our costs by 85% for the same performance as Sol.

[James Clear @JamesClear]

Remain playful as your responsibilities increase. It's easy to become serious when people and results depend on you, but nearly everyone's performance improves when they proceed lightly through the world.

[Aditya Jha @36Jha]

Holy chart crime

[Aditya Jha @36Jha]

Switching from Sol -> Terra for a task really made me feel the difference in speed. Remember seeing the TPS numbers somewhere, can't find it now. But I'd put it around a 2x difference.

[Aditya Jha @36Jha]

Databricks raising Series M. Stripe considering PayPal buyout. Is this not a failure of our financial markets that startups are just choosing not to IPO anymore? Startups see going public as an administrative mess, a mess that public markets have imposed upon companies to keep the investors "safe." At what point do we start to wonder, what are we saving the investors from. By keeping them inside a safe guarden are we locking them out of investment opportunities. Markets are a game of risk and reward, every participant must understand this, and be allowed to participate in what makes capitalism awesome.

[Aditya Jha @36Jha]

@franktisellano asked it to fix formatting of some formulas in a doc. would expect an agent to be able to do it, rather than giving me the steps for it.

[Aditya Jha @36Jha]

How is google docs gemini integration so bad!

[Itoflow @itoflowai]

Introducing symbol baseline cards.
While the agent works on an analysis, it will put up an agent card for you to get all the essential facts about a symbol. This will then ensure that the agent spends its time on going deeper into the harder questions rather than restating the known facts about a symbol.

[Aditya Jha @36Jha]

Is it weird I find frontend design the most unsolved piece of AI, because of the lack of originality, creativity and style on display there. Okay, I shouldn’t say unsolved, but you know what I mean.

[Itoflow @itoflowai]

AI power is not just a demand story now.

US regulators are speeding AI grid hookups, but data centers may have to pay upgrade costs or bring power.

The watch moves to GE Vernova ($GEV), Vistra and grid gear - not just utilities.

[Aditya Jha @36Jha]

Boring >> Exciting

[Itoflow @itoflowai]

OpenAI waiting on its IPO is the AI market's new tell. Suppliers can still win, but fresh AI listings are being marked harder. If $CBRS keeps fading while memory holds, this is funding stress - not AI demand failure.

[Aditya Jha @36Jha]

In a weird way, it kinda feels better building on top of a less smart model like gpt-5.4-mini vs 5.5. Simply because it exposes so many flaws in the design of your harness. Smarter models by their sheer brilliance paper over the issues, which shows up as token burn - and those are the issues which when resolved, make your smarter model go brr..
So kids, design with mini and run with regular

[Itoflow @itoflowai]

India's small-cap rally is the non-AI tape to watch.

SIQ's Small Cap Discovery is +25% since Mar 31 and still holding.

If small caps keep beating Nifty, breadth is real. If exits gap, size was the risk.

[Aditya Jha @36Jha]

@OpenAI @thsottiaux
found this annoying bug in codex app that when I have a subagent running and I double-esc interrupt the main agent, the subagent gets blocked too.

[Itoflow @itoflowai]

Behind every SIQ answer is a structured research workflow.

The agent defines the objective, selects the right data and tools, chooses the research methods, sets assumptions and bounds, then verifies the output.

For investment research, the workflow matters as much as the model.