2024 was a fantastic year for stocks but a terrible time for the tech industry. Last year I wrote about how layoffs were driven by companies’ needs to control costs while the tech industry has gotten more and more crowded with more and more engineers competing for fewer and fewer positions.

People clung onto the delusional hope that things would change and that the tech industry would get back to what it used to be in the zero interest rate era of the 2010’s.

I have been ringing the bell for 5 years saying otherwise; that jobs would be sent overseas and people would be replaced by contractors and that the gravy train couldn’t last forever. Anyone who has heeded my warning is in probably a solid position. Anyone who hasn’t has been taken out by the tide; laid off, unable to find work. To remain relevant, they’ve become the modern pamphleteer. A newsletter writer, sharing “insights” and “scoops” but fail to turn the information into anything meaningful or actionable. They are little less than content creators while decrying themselves as do-gooder journalists without any accountability for what they write.

It is tragic and frustrating to make these predictions while the rest of the industry marches to the beat of the drum of the latest hype, whether that is AI or coping with the latest round of layoffs by clinging to false hope. To me, the AI hype has been little more than a thin cover for just laying people off and scaring the masses. The state of AI is nowhere near ready. It is a fantastic assistant for writing cheap code snippets but not enough to reason and think about complex adaptive systems. Yet the press has run with the narrative.

But in the spirit of Warren Buffett, I too shall “praise by name, criticize by category.”

Forget the distractions. The only thing you need to do is be useful and uniquely so.

Business

I have not run any consulting or content creation this year. In fact, I actively chose not to because hiring was just too damn difficult. Many companies had opened headcounts but they were mostly backfills. I felt I would have been wasting my time.

In full transparency, allow me to outline what my fee structure is and put a number to my efforts. My standard fee is 15% for the first year. Let’s use Meta’s average salary as a high water mark since their salary is far above average. This does not account for people who may quit on me half way through and so on.

	Meta	Time to Prepare	Optimistic Fee / Hourly Wage	Actualized Hourly Wage
Junior	$181k	16 weeks, 64 hours	$27k/$421	$84
Midlevel	$327k	12 weeks, 52 hours	$49k/$942	$204
Senior	$500k	16 weeks, 64 hours	$75k/$1171	$234

These numbers are much higher than they were three years ago. The highest offer I secured for someone at Google back then was around $151k at the L3 level. In 2023, the highest offer I got for a new graduate client was around $120k at Microsoft in Canada.

You may have noticed that training junior and senior engineers takes longer than training mid-level engineers. This is because of how engineers are graded in technical interviews and the level of exposure they’ve had to real-world projects. Junior engineers are often new graduates, so I spend more time helping them understand the work ethic I expect, the level of detail required, and the time commitment involved. Despite the fact my guide exists and that I’ve clearly proven otherwise, I’m constantly battling the misinformation about work-life balance and the “just Leetcode bro” mentality.

If you don’t come to me already understanding what I stand for, don’t ask me for help. I’ve written out step-by-step instructions on exactly what you need to do. But most of the people who ask me for help have not read it or barely even tried to execute on it. Those who do this and ask to be my client are an automatic reject for me.

If the job market were like it was in 2021, I wouldn’t be excluding junior engineers. But right now, the market is too competitive. I covered this in last year’s letter where there are far too many junior engineers for too few spots available. Therefore, there is a very slim margin for mistakes. That part is not a problem: my methodology does not involve spamming interviews or gambling. It’s fairly consistent. But even if I were to train a junior engineer for 4 hours a week for 4 months, it’s unlikely we’d secure enough interviews, especially not with FAANG companies.

Mid-level engineers can often get by in interviews by nailing the coding portion and doing a mediocre job in system design. This is easier to train for because they’ve already been coding in production and are usually working with more experienced engineers who understand what it means to be a professional developer.

There are also very few L4 engineers looking to switch companies or move laterally in the current market. Everyone knows how tough things are, and most are just happy to hold on to their positions.

The Senior role is much more challenging. Most senior engineers with experience in complex, scalable systems don’t need my help. The ones who have never dealt with such systems likely lack a strong coding background, so they struggle in both coding and system design. If they do have solid coding skills, they’re likely L4 engineers trying to up-level to a senior role. I’ve helped with this before, but it’s tough to do consistently. I’m all about delivering consistent, real results.

How do you give someone an intuitive understanding of complex adaptive systems—how to focus on them in system design, and the inherent challenges—especially if they’ve never worked with one? This was made evident by my clients’ failure in 2023.

With such harsh filters, I am already unable to find more than 1 or 2 clients to coach per year. But that was back when I was actively creating content to find these clients. This is the real cost of the business: my time.

On top of the challenge of training engineers, there’s also the cost of producing content to attract new clients. If I want to appeal to more senior engineers, I’ll need to create higher-quality videos, and that takes a significant amount of time. In the past, I could produce one video a week that were little more than a rehash of my guide because the guide was written and the ideas were intuitive and familiar to me—principles I’ve internalized over my career.

But that attracts only a certain kind of client: a junior to midlevel engineer. Senior engineer clients usually require a higher bar and more interesting and nuanced content. But as the quality of the content I want to create improves, so does the time investment. I easily spend 16 hours a week planning, scripting, editing, and publishing each video. My Bazel video alone took a month of work, including editing, graphics creation, and testing to ensure it’s understandable for the audience. I can’t outsource this work to an editor because they wouldn’t have the technical expertise to grasp the details I’m explaining.

That is the video I am the most proud of because it encapsulates so much: engineering thinking, complexity, build system, system design, and it showcases my work. But that took such a long time to do I could maybe only do it 1 time a year.

When factoring in the expenses for producing content, I’ve also added a column to show the real wage, taking into account my clients who have failed. It’s never easy to see a client fail because, not only have I let them down, but I’ve also wasted months of my life. And on top of that, I have two cats who insist on distracting me all the time. Coaching and streaming with that level of stress doesn’t seem sustainable for me, especially since I have my own job that is very demanding.

At this point, I’m considering shifting my business model to a consultancy approach where I take a stake in someone’s salary, contingent on their promotion, but I haven’t made a final decision yet. If you’d like to discuss this further, feel free to reach out.

The Shadow of Jack Welch

To me, the writing is on the wall and I firmly predict that within the next 12 months, stricter RTO mandates will be enforced by tech companies.

Amazon has implemented a 5 days a week RTO policy.
Google’s Sergey Brin has written that he recommends being in the office at least every weekday.
Meta has a stricter RTO policy that will scan people for badge data where people who do not show up at least 3 times.
Box is pushing for its employees in the next 6 months to become full time in-office.
Apple is enforcing 3 days RTO as of Feb 1.

I have never considered hybrid nor remote work to be as effective as in-person because it took far too long to get answers from people with 1 day feedback loops the normal. This sentiment has been shared by Jamie Dimon, CEO of JP Morgan Chase. In fact, I moved because 2 years ago, I believed that RTO would be normalized.

Unlike the past decade of zero-interest rate policy where the power was with employees to job hop and negotiate for higher and higher pay, the power squarely in the hands of big tech and the employer. The employer can rank and yank at will. The employer can command more hours worked for less pay. Yes, the tech industry is still very well paid but if you don’t do what your manager asks, you will be replaced in 5 minutes.

It is off the back of this power change that I believe the RTO policy will be enforced. And this is just the beginning. We are seeing a repeat of the same practices that allowed Jack Welch, former CEO of GE, to thrive in the 1980s and 1990s.

His ruthless tactics were focused on efficiency and leaner operations, breaking the bureaucracy that had ruled GE. He would aggressively layoff the bottom 10% of performers with his famous 20-70-10 rule (20% A players, 70% B players, 10% C players), focus on operational efficiency, and financial engineering to manage earnings. He was so effective that over his 20 year tenure, he approximately eliminated 25% of the workforce and the stock price increased by 40x; a 20% compounded annual return.

What triggered such a response? His insane competitive instincts. With the manufacturing capacities of Asia in full force and the looming threat of globalization, General Electric was not able to keep up with Asia producing better goods at lower prices. To rectify this, Jack would layoff people who were not raising the bar instead of tenuring them. He believed it was better to be honest and upfront so that people would not cling to false hope and could be more effective elsewhere. Motivated by his childhood experiences and his mother and her eternal words “don’t kid yourself”, he was a very competitive CEO. In his biography, he quiped that

“Some think it’s cruel or brutal to remove the bottom 10 per cent of our people. It isn’t. It’s just the opposite. What I think is brutal and “false kindness” is keeping people around who aren’t going to grow and prosper.”

His logic: throughout your entire life, you are judged and graded. Why should that stop just because you got a job? He even went so far as to claim that if you fire fast, there will be jobs waiting for them. If you wait too long they'll be in the unemployment line because everyone will react too late to the bad news. By doing all this, Jack effectively broke the understanding that a person was to be at a company for life; that getting fired meant you did something so unforgivable the company had no choice.

Culturally, firing used to be seen as a tragic death of sorts in the past. Now, its just a bump in the road.

I believe we are about to repeat this paradigm in the tech industry. We should consider that one of the appeals of Google and the tech industry was a job for life: low to no risk of being fired unless absolutely necessary. In the book In The Plex, the decision to lay off in 2007/2008 in the depths of the crisis was not even founded on Google’s own needs

“While there was indeed an international financial meltdown and Google's growth had slowed, the company was in no serious danger… The worst of the financial world outside the googleplex created a great atmosphere to make tough decisions to cut waste.”

In 2008 and 2009, Google laid off 600 people out of 20,000 with 300 at Doubleclick and 200 in sales and marketing. At the time this represented 3% of the company in a workforce of 20,000.

But today, 12-25% of the company can be done simply at a whim. Last year, Google has focused heavily on trimming down its middle layers of management, with a 10% reduction in managers, directors, and vice presidents, part of an ongoing efficiency drive.

Meta, another tech giant, is facing similar scrutiny. With Mark Zuckerberg’s stated goal of making Meta “leaner” and more efficient, the company has already laid off thousands of employees, particularly in managerial and administrative roles cutting 5% in 2025 already.

Contrary to the past where layoffs were the big “no-no”, big tech doing aggressive layoffs of the bottom 5-10% of performers and prioritizing efficiency seems to become the norm like what Welch did with GE.

We can rule out financial stress for a few of these companies. Yes, many of these unprofitable tech companies have debts due in 2025 and a spike in their loan interest rates. But a cursory look at even healthy companies are doing layoffs still. I do not need to reiterate my 2024 analysis.

Company	1-Year Total Revenue Growth Rate	3-Year Total Revenue Growth Rate	1-Year Net Income Growth Rate	3-Year Net Income Growth Rate
Apple Inc	2.6	2.2	-4.7	-0.3
Microsoft Corp	15	13.4	12.4	12.9
Amazon.com Inc	11	10.7	94.7	21.1
Alphabet Inc	13.9	10.8	35.7	9.6
International Business Machines Corp	1.4	3	-19.7	1.6
DoorDash Inc	24.2	29.9	0	0
Snowflake Inc	29.2	43.8	0	-23.7
Snap Inc	16.4	9.2	0	-12.7
Lyft Inc	31.4	21.7	0	0

Furthermore, there is an obsession with using financial engineering tools like buybacks to boost the stock price. While buybacks are a legitimate tool to create value when the price is low, it is value destroying at exorbitantly expensive prices. With many companies at near all time highs, trading at nosebleed expensive multiples, companies have not slowed down on their buybacks.

Whatever the case is and why companies continually shout fear to the high heavens, the magic of the tech industry is no more. We’re becoming another mature corporate industry. History does rhyme and if the legacy of Jack Welch is any indication, I believe we will continue to see more and more restructures and operating management changes with top-down heavy directives.

Whether you think Welch is a blessing or curse on American industry, the fact of the matter is that there is a very good likelihood the tech industry will repeat the cycle. A fat decade of prosperity and excess may soon give way to a decade of efficiency and ruthless competition.

Employees should be fully prepared to obey or get fired. If you didn’t spend the last 5-10 years becoming skilled or valuable, you may be on your way out.

Or as Jack famously put it, be first or second in your market. If you can’t, either grow, fix, or sell it.

I personally choose to grow.

Better Engineers, Less Management

In the wake of massive layoffs and restructuring at tech giants like Google and Meta, many are questioning where we are headed in terms of workforce dynamics. The decision to cut back on middle management, as evidenced by Google's recent move to reduce management roles by 10%, is a growing trend that highlights the increasing push for corporate efficiency. This shift is not only about reducing headcount but also about recalibrating the organizational structure to streamline operations in a fast-evolving tech landscape.

While some of these managerial roles have been converted into non-managerial positions, others were eliminated entirely. This restructuring is indicative of the broader trend within the tech industry, where many companies are reassessing the value and necessity of expansive middle management layers. The tech sector, in particular, is increasingly favoring a flatter organizational structure that removes unnecessary bureaucracy and allows for faster decision-making.

The hope is that by cutting out layers of management, companies can focus more on innovation and product development, which are essential in staying competitive in the age of AI and automation. It reflects a deeper issue in the corporate world: bureaucracy creates more bureaucracy. The introduction of more managers, often meant to streamline decision-making, can instead lead to an ever-expanding web of oversight. Engineers and other technical staff often find themselves navigating through multiple layers of management, each layer adding its own set of demands, processes, and complications. What was intended to be a more efficient system often results in frustration, slowing down progress instead of speeding it up.

From a purely corporate perspective, a manager is there to deliver projects at larger and larger scopes, orchestrating people to deliver those projects, and moves to improve the efficiencies of those people from an organizational perspective. When you look at a manager, the only metric they are judged by is “did the team deliver what it promised?”. Everything else is just a supporting metric.

“They are also judged on the technical development and operations of things that impact multiple teams” Yeah but that’s all to deliver what they promised. It’s just that the team also improved things along the way.

You could group managers into a range from hands on to hands off with the technical lead manager the most hands on and an allocator the most hands off. A technical lead manager (TLM) may have 6-8 reports while an allocator may have 10-12 reports. Depending on what kind of manager you are, you may fall somewhere in between.

I think where the biggest issue lies is in the overabundance of allocators and not enough technicals. Let’s oversimplify a bit and focus on the allocator side because the TLM role is pretty self explanatory.

It may be easier to start by thinking of them as capital allocators and engineers are the businesses to be invested in. Engineers, like excess resources in a market, are critical for the innovation and growth of any organization. They possess the technical expertise necessary to develop products, systems, and solutions that, in turn, make money for their employer. Similarly, a business turns raw materials and intangibles into productive goods and services for the economy.

However, much like an oversupply of capital in a market, the sheer number of engineers without adequate direction can lead to wasted potential. Too many engineers working on different facets of a problem without clear coordination can result in duplicated efforts, inefficiency, and a lack of focus. In this way, engineers are analogous to untapped or underutilized resources that require careful management to maximize their impact.

This is where managers come in. They are tasked with directing and harnessing the potential of these engineers. Their role is to ensure that the collective efforts of the engineers align with broader organizational goals; to organize and guide this chaos to productive means. In effect, middle management are the layer that mediates between the top brass and visionaries and the people who just treat this as a 9-5 and don’t really care. Management indeed worked for Google both when Google was founded (per In the Plex, fantastic book) and in 2012 in a Harvard Business Review study. Even the best and brightest need some middle managers.

But any good thing can be taken too far. Manager effectiveness is often hampered by inefficiencies in management and the over-saturation of roles. An overabundance of managers can disrupt the focus of engineers, forcing them to chase short-term objectives or engage in activities that distract from the bigger picture.

This is not unlike financial markets: when there is too much money chasing too few opportunities, people reach for returns or participants take profits from each other in a zero sum game. Hedge funds, while designed to maximize returns, often overcomplicate the market by focusing on short-term gains and speculative strategies rather than long-term value creation. Not only do many managers do this, managers may engage in turf wars to stake claims on the gains produced by engineers instead of making the engineers more productive. An excess of managers can lead to diminishing returns where there are too many managers trying to allocate and orchestrate headcount and not enough heads to produce the work and value to be allocated.

Both scenarios underscore the importance of balance—too few managers or hedge funds might leave an organization or market under-optimized, while too many can stifle progress and productivity.

Continuing the parallel, if Charlie Munger laments the fact that there are too many people running hedge funds and trying to gamble on the price of American businesses, then we in the tech industry should lament the excess of managers trying to gamble their careers on the backs of engineers who, in most cases, don’t really need to be managed.

At most, the A player engineers who are highly productive really need managers to move political roadblocks. It is the B players and lower that need to be managed and overseen most of the time.

How do you know if you need to be managed? To summarize a long talk (and Jack Welch’s own thoughts) the rubric for A, B, and C players:

(Top 20%) A player: “Thank god you’re here!”
(Middle 70%) B player: “Oh you’re here.”
(Bottom 10%) C player: “Why are you here?”

Personally, I’d have more a larger percentage of people in C and fewer people in A.

If a manager simply winds up with a team that produces stellar results without his involvement, is he still a good manager? It may seem obvious to some people but upon a nuanced take, I don’t know the answer to this question: it takes a lot of courage to not touch something that is perfect.

Make sure you always work to be the kind of person people rely on when shit hits the fan. Managers dream of a team they don’t have to have any involvement in and reap the rewards. And you, as an engineer, should dream of a team where management doesn’t need to get involved. Of course, that means being the type of engineer and problem solver who doesn’t need to be managed either. A society of deserved trust is truly a wonderful utopia. But we’ll settle for a team.

Where Are The Subscribers’ Careers?

In today’s fast-moving digital landscape, newsletters have become a dominant way to consume news and insights. Whether focused on tech, culture, or business, they promise exclusive content delivered straight to our inboxes—often with urgency and importance. But what actually makes a newsletter valuable? And, more importantly, what are these newsletter writers really selling?

A closer look at the tech newsletter ecosystem reveals a troubling trend: much of the content is regurgitated, sensationalized, or driven by narratives that are not just misleading but actively harmful. These narratives shape careers, investment decisions, and even entire industries—often for the worse.

One of the biggest problems with tech newsletters today is their reliance on recycled content, particularly when covering major news events like layoffs. Rather than providing nuanced analysis, many simply restate “facts” without questioning the narrative. Worse, the analysis they do provide is often deeply flawed, reinforcing misunderstandings and making the industry collectively dumber.

This isn’t just criticism for the sake of it—I mean this in the most literal sense. There is no independent investigation, no deep dives into data. Instead, newsletter writers repeat what others have said, creating an echo chamber where bad takes go unchallenged. If the initial premise is wrong, the industry simply moves on to the next trending topic, leaving misinformation unchecked.

I highlighted this issue last year when newsletter writers collectively pushed the narrative that tax policies around depreciation vs. expensing investments were causing a cash crunch and triggering layoffs. But even a cursory glance at layoff data would have revealed that well-capitalized companies were cutting staff as well. The narrative didn’t align with the facts—but nobody bothered to check.

And yet, there are no consequences for getting it wrong. Newsletter consumers have alarmingly short memories, and as a result, bad analysis goes unpunished. Perhaps these writers have found the audience they deserve.

The quality of these newsletters can’t be sold on the merit of their information—because how could it be? A single person juggling a daily newsletter, a podcast, a YouTube channel, and social media updates has no time to ensure their reporting is accurate.

Instead, the credibility of these newsletters is often built on prestige. “Ex-FAANG” status is treated as a badge of authority, as if working at a big tech company automatically translates to journalistic integrity. This is no different from the wave of course sellers a decade ago—"Buy my course because I’m ex-FAANG" has simply evolved into "Trust my analysis because I’m ex-FAANG." But name recognition is not a substitute for rigorous analysis.

Bad information has real consequences. It shapes public discourse, influences investment decisions, and affects people’s careers. If you make decisions based on flawed analysis, you're more likely to make bad choices.

How many people sat paralyzed during the wave of tech layoffs, praying they wouldn’t be next, simply because they believed a misleading tax narrative? How many jumped into crypto and pushed Bitcoin as a hedge against the Federal Reserve—egged on by newsletter writers who couldn’t care less about the accuracy of their claims?

How many took their own lives because they followed an influencer’s promotion and lost everything? And no, don’t give me that “not financial advice” and “you’re responsible” bullshit. Magic words don’t excuse you from promoting life-ruining activities. If it was the case, then there would be no need for a FDA. Just buy whatever food you want and hope it isn’t poisoned. Buyer beware after all.

Perhaps basic human decency isn’t in these people’s vocabulary. I hope they pick up a dictionary.

Crypto scammer in Squid Game, Season 2. Do you really want to share the same excuses and reasoning as a scammer?

This happens to lesser degrees as well. Choose the wrong company or the wrong tech stack, and you could pay for it for the rest of your career. I don’t know anyone who proudly calls themselves an ex-crypto engineer—unless they’re still in space, have pivoted to another grift, or have no choice but to stay. The end of a career doesn’t necessarily mean the end of life. But it does mean significant hardship because all the experience and knowledge you’ve built up is effectively rendered useless.

Personally, I refuse to align my career with ethically questionable ventures or companies I wouldn’t be proud to have on my résumé. I want longevity in my career, not short-term hype which means focusing on areas of engineering where there will be a tangible and viable demand in the next few years. I have always done it this way; I have staked my career on Android for better or worse for the past 13 years. I’ve compounded that experience without having to throw away the past knowledge I’ve accumulated.

As the saying goes:

“Beware of an old man in a profession where men usually die young.”

How Did We Get Here?

Rather than taking another victory lap when yet another tech newsletter is exposed for lacking substance, let’s put this issue in a broader historical context. Allow me to introduce you to John Law.

Law was an 18th-century economist, a convicted murderer, a notorious womanizer—and a pamphleteer with radical ideas about banking. He proposed the creation of a new financial institution, the Banque Royale, which had the power to print paper money that could be exchanged at par with French state currency. The bank’s value was tied to the Mississippi Company, a venture Law created that based its worth on speculative future profits from French land holdings in North America.

Speculators flocked to the Mississippi Company, using it—and by extension, Law’s bank—as a tool for gambling. Eager to see his theories play out, Law continuously propped up the stock price by purchasing shares at ever-increasing valuations. He even allowed French government debts to be converted into company shares. All these operations were fueled by newly issued paper money, with the assumption that as long as no one redeemed their notes en masse, the system would hold.

Until it didn’t.

Doubt in the scheme’s viability led to a bank run, as people rushed to redeem their notes. Law fled, the French government was forced to intervene and assume the debts, and the resulting financial crisis contributed to the French Revolution.

This episode of financial history eerily mirrors the utopian vision that crypto and Bitcoin enthusiasts have championed—and continue to champion. The rise of alternative, unregulated financial institutions has repeatedly led to disaster. The idea that financial systems should operate outside of government regulation is not some groundbreaking innovation; it's a recycled bad idea that history has disproven time and again.

Some might argue that John Law’s experiment ultimately led to positive change, such as the emergence of French democracy. But that’s simply hindsight bias. The immediate reality was economic devastation.

The larger takeaway here is simple: a handful of people with reckless and poorly conceived ideas can ruin countless lives. The crypto boom of recent years is just another chapter in this recurring saga, and I have yet to meet anyone who is genuinely proud of having spent years working in crypto.

Crypto wasn’t just pushed by influencers—it was aggressively promoted by tech influencers, who positioned it as “the future” that would “democratize” and “decentralize” finance. Web3, blockchain, and other buzzwords were sold as revolutionary concepts, despite the fact that no open-source software project has ever succeeded without centralized leadership. Even the most successful open-source project, Linux, relies on Linus Torvalds at the helm.

The shift from traditional journalism to content-driven newsletters has made this problem worse. Writers today are incentivized by clicks and subscriptions, not by the pursuit of truth. This creates a modern parallel to 18th-century pamphleteers, who were often criticized for spreading sensationalized, unverified information to provoke emotional reactions for political or personal gain.

The tech industry, in particular, has become a breeding ground for misinformation, oversimplified opinions, and half-baked theories masquerading as expert analysis. Nowhere is this more dangerous than in discussions about AI. The sheer volume of uninformed takes drowns out meaningful discourse, polluting an incredibly important conversation with hype, fear-mongering, and blatant misunderstandings.

Tech newsletters today don’t just misinform—they actively shape narratives that influence careers, investments, and industry trends. But if history has taught us anything, it’s that bad ideas, left unchecked, have real consequences. And as we’ve seen before, when speculation and misinformation run rampant, it’s not just the gamblers who lose—it’s everyone.

While it may have happened in the crypto space in the past, it is happening again as we enter the AI space. I was happy to ignore this conversation because crypto was purely for gamblers and speculators that, if imploded, would have no real effect on the tech industry. But noise around the discourse for AI has reached a fevered pitch that I think needs to be addressed because this attitude is beginning to affect the entire tech industry, both the careers of the engineers and where investors choose to put their money.

On the career side, I need not repeat my criticisms of how engineers have wasted their careers by just being code monkeys. Time has already started to expose these individuals and their careers will go out with a whimper (just look who’s just selling information and influence versus producing real world results). It’s one thing for someone to do it at the end of their career in retirement. It’s another to be a sophomore acting like a boss lord.

Euphoria

Kill Bill Scene - How the main character can't tell the truth to herself about her target

Who knew that the COVID hiring bubble was really the last big tech job bull market right before it all died? Did nobody really see that the entire job market was kidding itself for the past 5-7 years about big tech’s glorious and frothy future?

Maybe the scammers and influencers were onto something and decided to pull one last big heist before disappearing and spawned an entire industry dedicated to spreading lies and misinformation and milking their titles for one last payday.

But I want to close out this post with some hope. AI will definitely become useful but it is too soon to say whether or not which models and technologies will stay. Which ways of training models are truly effective? What data sets are valuable and what aren’t? These are questions that are still being answered. It is a greenfield with ripe but shrinking opportunities. If the past is any indication, shakeouts in new paradigms occur every 5 years or so and we’re only 2-3 years into the AI hype cycle. It is too soon to call a winner.

As for me, I choose to focus on the technologies that I truly understand and work to eliminate that ignorance every day bit by bit through competence. In my belief, no matter what happens, people will still need to have a phone in their pocket to communicate and take selfies with; whether an AI program is doing it for them or not. While I am studying these whitepapers, I just don’t have a good grasp of them to truly understand what goes on under the hood.

But while writing this letter, I was horrified to see that there were people with no real knowledge of AI who were creating tools to act dishonestly and who created these tools with dishonest means with VC money backing. As one bubble pops, another one forms.

When this will stop, I have no idea. I tread cautiously and rationally.

SOTU 2025: Be Useful