Biren Shah

@ShearedSheeple @TheStalwart Giving computers the ability to hallucinate is going to change the world. But the hallucinations have to be harnessed and controlled. That work is still being done. Like most new tech, marketing and expectations jump past the years of experimentation that deliver on the promise

@biren215 @TheStalwart They're applying the tech in the exact way it was marketed to them If investors thought it was just a fun tool that can (at best) supplement real expertise, then there would be no AI bubble

Something to grasp: the hallucinations are the feature, not the bug. LLMs are amazing *because* they hallucinate. Now code can do something it wasn’t specifically told to do. It doesn’t replace deterministic code. It’s a new special function you can call that didn’t exist before.

@D_M_Ventures @TheStalwart Gasoline analogy: You don't want gasoline all the time, but when you do want it, it's *because* it's very flammable and energy dense. Those are the features. Just like sharpness isn't a bug in knife. Bugs are not strengths misapplied. Bugs are bugs.

@biren215 @TheStalwart I don’t disagree, but saying it is a feature when it isn’t for that usecase doesn’t mean anything. It seems great a coding, but I don’t code, and it’s not good at my job at all. That’s my point.

@indiemusicfan4 @TheStalwart Totally fair. Hard for me to say what the downsides of the other way would be--since it costs like $100m+ to find out. Maybe it'd just be poor engagement (who wants to hear IDK for the 100th time?) Maybe LLMs only work when IDK isn't an option. They can be damn lazy as it is.

@biren215 @TheStalwart hallucinations happen because models are trained and rewarded to provide answers, not say "I don't know". but "I don't know" is often a better answer than just lying

@D_M_Ventures @TheStalwart LLMs weren’t made to be junior lawyers or junior engineers or anything else. They’re just an interesting extension of some really complex math. The fact that they’re 90% l useful for so many things is honestly shocking but also not good enough (and people are working on that)

@TheStalwart @biren215 Yeah I’ve no idea what he is talking about. As a lawyer, when you ask it to review documents and it just makes stuff up, that is definitely not a feature for me. That’s the opposite!

@TheStalwart It’s not that different from people with big imaginations. It’s easy to lose touch with reality. The thing you’re seeing is people trying to apply the tech carelessly. It’s like gasoline has a “bug” in being so flammable because people set themselves on fire sometimes.

@biren215 To the extent that people use LLMs for information, it's a bug

@LEGOLAND_CA that’s two restaurants that can’t open at their posted time. What kind of shop are you running over here?

Spent the last day trying out GPT-5.4 for where I normally use Opus 4.6. The biggest difference I see is that that GPT-5.4 is both more strict about workflow and WAY less able to conform to specs. It's half the price, but probably not worth the slowed development cycles.

This is a terrible chart crime. 24 months ago was early 2024, or like 6 months after AI "became a thing". Cohorts that have aged enough to generate 24-month retention data are: 1. small 2. made of the earliest of early adopters, i.e. "the golden cohort" This is a common PM lie

I've been working with AI to build a system to write real software (not one-shot demos, software that's actually maintainable). I can say that AGI is far, far away AND that the coming revolution for knowledge workers is going to be dizzying. The hype is too hyped though.

@MapleStax Trade!

I’ve made over $100,000 in just 3 months using my custom CBC strategy. Now I turned it into a FREE TradingView indicator that tells you exactly when to buy and sell—no guesswork. I’ve been testing it for months, and it’s been printing. 📈 Like + Comment “Trade” - I’ll DM it to you. (Must be following to DM) $SPY $SPX

@simranpatel___ Sure. I followed you back, but I can dm you. I have no idea how the new chat permissions work…

@biren215 If you don't mind, you can message me directly. I have followed you, then we can know about each other.

The most limiting thing I see with vibe coding is that that LLMs are absolutely terrible at refactoring. There's just no way for the LLM to read something and not treat it as "fact". Refactoring is all about here's what is and I don't want it to be this way. LLMs suck at that

@simranpatel___ I started building some side projects in June of last year and just had to keep evolving as I fell into one trap after another lol

@biren215 That’s a strong mindset shift designing for ephemerality plus blast walls feels very AI-era native. What helped you decide where to draw those boundaries?

@simranpatel___ I’m sure a lot of this I just some form of best practice for some coding style out there—but it was a big shift for me

@biren215 That kind of reset is tough but powerful what shift in your mental model has made the biggest difference so far?

@adamtaggart He doesn’t share how he does it, but @DaveHcontrarian has been amazing at calling price targets. Totally worth an in depth interview.

I get solicited hourly, literally, by talking heads (or their PR agents) who want to appear on my show I pass on a LOT of them. Most are just not a good fit or high enough caliber for the Thoughtful Money audience Who would YOU most like to see me interview in 2026?

A thought: runaway corporatism has wounded our nation. The only “cure” being offered/pursued is an abridgment of property rights (cure way worse than the disease). But the idea that the only stakeholder that matters is a law of nature. Maybe there’s a better way in there

If anything, my parents and in-laws would tell you how much better America is than India (and up until recently, Britain too). We’re the first in line to tell anyone who’d listen to not “import the 3rd world”. It’s brutal out there, and “Indian” spans a huge range of culture and education.

@catpoopburglar @drishtadyumn I’m pretty much the exact group you’re talking about, and I have never in my life met any Indian who hates the British or has ever talked about colonial resentment. My dad talked about how much the Brits extracted from India, but in his next breath would compare it the Moguls

Peter Thiel's observation about deep corruption in India isn't incorrect, but the issue is his claim that Indian society is steeped in anti-Western or anti-colonial hostility. That narrative is invented by people looking for a rationale to justify their pre-existing contempt toward Indians. This reveals more about his own prejudices than Indian reality. The anger visible in India targets two groups: Islamist extremists and the country’s own parasitic political elite. It is not aimed at Western culture. There is no widespread obsession with hating colonizers. That fixation belongs to black people in the West, not Indians. The irony is that Indian society, for all its faults, has largely made its peace with the British legacy. Should we be aware of colonial atrocities? Yes. Should we actively seethe about it? No. That would be impractical and a waste of energy better directed at present-day survival and advancement. Thiel is copying an American script. In that script, every non-white group is assumed to carry racial anger, no matter the facts. He takes that idea, swaps in Indians, and repeats it. The result is a caricature, not an insight. The real resentment sits with the people who need to believe others hate them. They project their own imagined grievances onto Indians so they can pose as victims. What makes Thiel's misread especially telling is its political context. The MAGA crazies genuinely believes brown immigrants have racial grievances against whites. They need to believe it because it validates their own resentment and frustrations. In reality, Indians were extremely empathetic toward the violence and harm they experienced from blacks and Muslims. They deliberately miss how Indians have often been more receptive to Islamic atrocities and rape gangs across Europe than many white Americans. This reflects the gap between Thiel's projection and ground realities. After years of reading his books, essays, and interviews, this is where I draw the line. He's not just wrong; he is being intellectually lazy. This signals either a deterioration of his judgment skills or a surrender to base tribal instincts.