I exactly have been on same thoughts for last 2-3 months.. it has just become difficult to trust high volume moves.. Regarding AEROFLEX today, i thought it was perfect liquidity day for big funds or investors to decrease AI exposure.. but delivery volume so less means probably I am wrong.
But overall everywhere liquidity being sucked out, rotations can become more fast n sharp
Regards to index, why do you it’s only smallcap 50 index.. I thought smallcap 250 index has more funds tracking
The point is how to navigate this situations?
Huge false moves will stop out any technofunda person and even though earnings are good how practical it is to hold the positions for a month or two.
Face of the markets have changed considerably over last 2-3 years, due to tech and access to easy information.
This democratization has killed YoY. Now it’s only QoQ. A great YoY is a negative if the QoQ is lukewarm and the only time YoY matters is when it looks bad. Then it takes the stock down.
As a result, there is zero sympathy or patience for cyclicality or seasonality of a business.
None of this affects long haul players, but more volatility. And more opportunities to buy at low.
Doesn’t help that the markets have been falling for 2 years, so majority of players would be swinging in and out.
It’s not a T-20, it’s Test Cricket - Every body skill, patience, time and money will be tested.
It takes years, if not, at least a decade, to understand the art of investing.
Keep it simple - Invest in what u study, learn, understand and finally conviction - what ever be the style - whether momentum or value or anything - “Circle of Competence”
That’s why impatient investors hand over the wealth to patient investors in the market.
U can make quick money in a stock or in a trend or two, only for a short period - but to “sustain money/wealth” it over a period of time, is it possible ?? - As there is underlying business in the real world behind the stock name, which is changing dynamically as the times we live in.
This is “ THE GAME “ or “ART” one has to master - Only way is through “ TIME” in the Markets through thick and thin.
@narayans79 - There is a complete overlap between the 50 and 250. For someone managing smallcap indices, picking the 3-4 stocks from the 50 makes the most sense to make an impact. You can see how the perf diminishes as you go to the larger indices like the 100 or 250 though same stocks are present but at lower weightage. A lot of algorithms track these indices to take trades so as one index turns it attracts algos to increase risk, as this acts like an early signal for market health.
@Nakuljain04 - Volatility is hard. The best way I know is to move towards cash or more liquid names when I expect volatility and significant short-term drawdowns. Last year as well, this is what led me to buy a whole lot of Sai life in sept and then Hind zinc later on in December before moving on to the AI trades in Feb. Avoiding these steep drawdowns is crucial for maintaining a good alpha.
I know people who are undeterred by this kind of volatility as well and go through it because it suits their personality. I am personally OK with taking business risk or sectoral risk but always averse to market risk, so I do what I do. It may not suit everyone and there is no right way. If I could cultivate the ability take a 30% drawdown at pf level, maybe I would be able to zen through it but so far the max dd I have ever had is 17% and this level I have seen 3-4 times in the last few yrs as invariably the market bounces for me from this level but this time I felt it may not and I did not even want to face a 17% dd considering in absolute amounts, it was too big due to pf performance in the last year or so. I still have sufficient amount invested to not feel left out if there is a strong pullback as it happened this week.
@Garuna - Yes, the sideways market means most people have had zero or negative returns in the last 2 years and so are becoming more and more short-term to make something in the market. Its a zero/negative sum game at present and I think no one can make consistent money with these strategies though it feels lucrative. I have backtested earnings based strategies and they do really well but its so easy to build these with claude that I am certain there a few hundred of these bots trading the market which will explode to a few thousand shortly. There is no free lunch and these will find it hard to churn good returns. Right now though, they are doing quite well and its coming from technofunda/funda guys returns and possibly mental health, considering the volatility.
I have been thinking about the impact of AI on us through several posts here and this is a continuation of the same. I read a beautiful work on this (Introduction | The Last Economy) which I highly recommend. One of the core ideas is that intelligence so far has been priced as scarce by our economy but it is becoming so abundant that it will register as a collapse. Essentially people wondering why AI is not adding to GDP growth are looking at the wrong metric or rather, looking at the only metric we know to look at.
I will give a few examples. I was able to build a guitar amp sim in few days and I use it over the amp sim I paid a lot of money for. I have made the source open and the app free to use. This sort of devaluation of IP will come as a deluge to anything that is a software product or in general, knowledge work. This is great for me as a user as it adds tremendous value to me but will register as degrowth for the company making amp sims. We had a serious trouble with our 3D printer which without AI, we may have had to throw out. Claude helped my son debug the issue exactly to the thermistor which was burnt. This means we don’t buy another printer which will register as bad in the GDP but it is great for me. How many such great things are going to be measured as bad because we are measuring the wrong thing?
I have speculated here in the past that we might see token consumption being added to GDP. I think some form of it will most likely happen, though token is not really a good measure of intelligence since not all models need the same amount of tokens to solve the same problem. The argument I am making is that the revenue of Anthropic/OpenAI (and other AI companies whose revenue comes from selling tokens) is not a good measure of value added by them as the true value is far higher but currently not measurable or as the book argues, a single cured cancer knocks off $400k from GDP but saves a life.
The other thing is where revenue for AI companies will come from. The trillions in revenue these companies are vying for can only from from current wages. The world GDP is $120t of which $60t or so apparently is wages. The AI companies are eyeing 5-10% of this in the next 5+ yrs. This can only come through layoffs and this is already underway with several tech companies starting them to fund their AI capex (the Microsoft one being the latest). This can only mean severe loss off capacity to survive/consume which should hit the GDP significantly in the medium term if left unchecked.
Since the economy is a dynamic thing, all this may act as a balancing force. Without GDP growth, we might see funding environment for AI get harder and harder, through both debt and equity and also internal accruals (like Google struggling with search) which means we don’t get AI tokens cheap enough to matter. This doesn’t mean Indian IT will do well - that ship has sailed. It just means the adoption may take more time as the economy adjusts to this new reality. We can see one aspect of this (balancing force) already in memory prices which have gone up so much to make returns relatively unviable. This will give everyone time to adapt and go through this phase transition which at this point seems like a likely outcome. So though there is serious demand for tokens, supply might be hard to put up because of various constraints from cost of capital to memory prices to political will (may happen if democrats take both house and senate in the mid-terms). I don’t think it changes the value of companies participating in the DC infra much (just changes the shape of cash flows) but over a longer period other things can change as well (like changes in technology and competition)
Just thinking aloud
Do you see a form of Universal Basic Income, sort of global Ladki Behna Yojana coming into picture because a lot of wealth is generated by the consumption economy by the wage earners, and those who do not earn, do not spend. An important realisation I have had in recent times is that when it comes to spending money, assets matter way less than cash flow. People could have FIRE level assets but if their cash flow stops due to job loss, the mentality changes from abundance to scarcity and consequently spending over things they do not need. The alternative way of boosting spending could be the AI companies themselves trying to put money in the pockets of the consumers either directly or indirectly via couponisation or monetisation of used token bundles because the methods of GDP slowdown you’ve described will definitely be accompanied by a consumption slowdown as the world factors in the reduced need of human intelligence and consequent job loss. Americans traditionally have been very good at making people buy things - they will have to figure out a way.
I can be wrong/dumb here, so feel free to remove this post if it’s irrelevant.
I’ve been seeing a lot of fear about AI, and it’s understandable. Models are getting cheaper, and we now get fairly good ones for free. But with so many users, now and in the future, compute will become a bottleneck, and huge amounts are being invested in infrastructure. This has created a vicious cycle: the shovel providers invest in AI labs, and the labs spend that money on their products, which flows back into the providers’ revenue. We don’t know who will win. I don’t see the labs becoming profitable anytime soon, at least not until enough people pay to justify their huge valuations.
Reading these blogs made me feel miserable, wondering what will still be valuable. As you said apps you build with AI improve their own lives but don’t add to GDP. But apps which solves people’s problems can be great but I’m skeptical here because of the number of duopolies I’m seeing right now, and I fear about “entering right time would make you successful” when it is very easy to get competitors replicate what you’re doing.
I’m optimistic about experiences as an industry, so here’s my question: if productivity increases many times over, will people seek out new experiences and be willing to spend a lot on them? Or, if AI disrupts jobs on a large scale, will fear keep people at home, making experiences less attractive than cheaper alternatives like YouTube or OTT?
My guess is that it could be both, in a “K-shaped” split. I think this matters a lot for India. Our consuming class is already small, only around 130–140 million people, and the jobs most exposed to AI (IT services, BPO, back-office work) are the ones that built our urban middle class. If those jobs shrink, the group that buys flats, flights and concert tickets could split, and the path up for tier-2 and tier-3 graduates could narrow.
Could AI-driven job losses in India’s IT and services sector shrink the experience economy, or will growth at the top make up for it?
Also, do you think AI will make healthcare and education less costlier than it is today?
This is not sustainable. Will lead to social unrest. We are already seeing glimpses of the same around the world.
Other important issue is population collapse happening in economically well off pockets.
Dear @phreakv6 bharani, thank you so much for keep sharing your thoughts. Really appreciate the clarity with which you explain things. Going through your posts is always a great learning experience.
I am wondering if you could share a bit about how you use AI in your investing, especially for building tools or programs to help with your research. Would love to understand your overall approach using AI. What kind of things you use AI for and may a short not on how you achieved it usig IA, what parts of the process you have automated, how you extract and analyse information, and how you eventually use it to build your watchlist.
I’m sure it would be very useful for many of us who are exploring ways to use AI in our investing journey. Looking forward for your kind response.
AI trade is unwinding across world now, after democrat win, market expect lot of policy changes on AI front which may slowdown growth in a bubble sector, if anthropic IPO fails, its a bloodbath.
This has played out this week. Everything looks emotional and algorithmic right now with sustained selling. AI trades looked like the sure thing so there was lot of leverage and unlike the quick bounce since Leopold went bust in June, this time the chaos is more widespread. However, I think we should look past all the noise into reality and probabilities. The kneejerk reaction will be done this week and reality will hit next week. Can the US risk not doing AI buildouts at current accelerated pace? What does it do to their economy (a recession will be certain) and their debt?
I think we should see what’s possible and what’s not possible. I think current models will get served and I dont see a slowdown in regulation on superintelligence stopping the serving of Opus 5.5 - so inference is safe. Training new models might face hurdles with committees and boards and red tape and someone’s neck on the line if things go wrong. If buildouts slow down in the US, we will have direct impact on companies like TD/Mtar/Stltech which are US facing plays. There is a possibility that these may move out to the gulf or to India if we are lucky and do what is required (power, quicker approvals). I don’t think this puts India facing plays like Aeroflex, Vivid, Marine etc. under great risk fundamentally. Sentiment-wise as I was thinking in my last few posts, we may not see 100 p/e in these names again. They may settle between 60-80 at most.
I do however think there will a second-coming for AI plays, just like the Covid second wave or Bitcoin’s rally in 2020/21 and 70% drawdown in 2022 and a 6x move in 2023/24. As of now, the noise is deafening that AI is dead which is an extreme stand to take. Just look out for consolidation and good valuations and try not to get too pessimistic just becase the price is asking you to be so.
I continue to remain 40%+ cash personally. I had sold TD/Mtar over last few months and am completely out of them. I had part sold Aeroflex as well and have a small position in Vivid too (both together make up ~10% of pf). I will most likely reinvest back in AI names which I understand well, rather than chase financials which I have always dreaded owning.
As of pf earnings, everything has been great so far just like last qtr with the last one coming in yesterday (Sai).
The best so far which surprised me on the upside is Yash and Aeroflex. The rest were as per expectations. On response to earnings though everything else has been tepid/poor after a week or two (Sai may end up the same way in 2 weeks though today’s perf is encouraging). Yash is the only one which looks like it might be able to take out earnings day highs.
Disc: I am a novice and share my opinions on my pf and sectors I invest in. I could be biased due to my positioning (40%+ cash)
@phreakv6 Thank you for your insights and balanced commentary in these turbulent times. Despite all the noise, I agree that there could be massive tailwinds coming for the Indian players like Aeroflex and others in the data center space. How do you view the change in depreciation method adopted by Aeroflex?
Hi Manohar
Earnings reaction to Venus Remedies was brutal despite decent number.
Any change of view post results?
Disc. Holding 6% of PF
