DesaiAshu

The Department of Defense was named as such after the detonations of the atomic bomb in Hiroshima and Nagasaki

We - as a humanity - collectively recognized the weight of our creation, and decided to walk back

Discussing “AI alignment” in the same breadth as aligning with a “Department of War” (in any country) is simply not an intellectually sound position

None of the countries we’ve attacked this year pose an existential threat to humanity. In contrast, striking first and pulling Europe, Russia, and China into a hot war beginning in the Middle East surely poses a greater collective threat than bioweapons, sentient AI, or the other typical “AI alignment” concerns

Why aren’t there more dissidents among the researcher ranks?

It's not "just" a $200m contract, it's the start of a lucrative relationship

1. Stargate seemed to require a dedicated press conference by the President to achieve funding targets. Why risk that level of politicization if it didn't?

2. Greg Brockman donated $25mil to Trump MAGA Super PAC last year. Why risk so much political backlash for a low leverage return of $200m on $25m spent?

3. During WW2, military spend shot from 2% to 40% of GDP. The administration is requesting $1.5T military budget for FY2027, up from $0.8T for FY2025. They have made clear in the past 2 months that they plan to use it and are not stopping anytime soon

If you believe "software eats the world" it is reasonable to expect the share of total military spend to be captured by software companies to increase dramatically over the next decade. $100B (10% of capture) is a reasonable possibility for domestic military AI TAM in FY2027 if the spending increase is approved (so far, Republicans have not broken rank with the administration on any meaningful policy)

If US military actions continue to accelerate, other countries will also ratchet up military spend - largely on nuclear arsenals and AI drones (France already announced increase of their arsenal). This further increases the addressable TAM

Given the competition and lack of moat in the consumer/enterprise markets, I am not sure that there is a viable path for OpenAI to cover it's losses and fund it's infrastructure ambitions without becoming the preferred AI vendor for a rapidly increasing military budget. The devices bet seems to be the most practical alternative, but there is far more competition both domestically (Apple, Google, Motorola) and globally (Xiaomi, Samsung, Huawei) than there is for military AI

Having run an unprofitable P&L for a decade, I can confidently state that a healthy balance sheet is the only way to maintain and defend one's core values and principles. As the "alignment" folks on the AI industry are likely to learn - the road to hell (aka a heavily militarized world) is oft paved with the best intentions

Basically all competitive sports in the US work like this

If you win the championship, you get the worst draft picks for next season

Do you believe they discriminate against winning teams and reduced the quality of the sport? The Yankees definitely complained a lot about it

Spending money to give scholarships to people who are coming out of 300 years of tariff imposed poverty to access the same education as those who can easily afford to pay their food/housing costs in college is "the dumbest idea we have ever come up with" ?

Please recall we paid more in reparations to Germany post WW2 than we paid to India post-colonialism

We seem to not have much problem undo'ing the Nazis' wrongs with our money, why do we have a problem uplifting the Nigerians?

You're not able to publish cutting edge research in an era where you have LLMs and Arxiv?

Academia seems more open and competitive today than ever before, with more weight and influence given to more universities around the world

Adjacent medicines have seen major improvements: eg Ketamine was a significant improvement from PCP (notably, less psychosis and safe enough to use off the battlefield / with children)

“Removing the worst and most fatal danger” is a laudable goal with Fentanyl given the absurd rate of ODs

It is generally understood in the industry that around half of universities are in significant debt / financial distress (started prior to Covid // the demographic peak // recent DoE cuts). Graduate underemployment is also quite high due to a lack of alignment (or perhaps slow alignment) of degree programs to career outcomes

https://www.philadelphiafed.org/-/media/frbp/assets/working-...

https://www.stlouisfed.org/open-vault/2025/aug/jobs-degrees-...

Ideas for solutions here:

https://files.eric.ed.gov/fulltext/ED604299.pdf

It would require building public transit and higher density residential housing over the next two boom/bust cycles

Stanford has a $40 billion endowment for 8k undergrads. UCLA has a $10 billion endowment for 34k undergrads. Naturally, the class sizes will be much larger. The UC system does not put 100% of students at UC Berkeley and UCLA, they distribute it across several campuses and distance education and maintain a leveling system that helps promising research talent be in the room with experienced researchers

Despite rising costs, a college degree is still a positive lifetime investment for students (not to mention the positive externalities educated populations have on society at large). The bulk of US college students attend colleges who do not have the resources to build high-quality, industry relevant curriculum, train teachers to teach with modern pedagogy, and efficiently manage dorms, student affairs, and other administrative infrastructure

HYPSM choosing to share land, curriculum, expertise, and administrative infrastructure through network'd partnerships would lead to massive economies of scale and a broad reduction of educational costs. Another way to think about this - is one city of 1 million people more efficient to run per capita than 10 cities of 100k people? The answer is a resounding yes due to urban scaling. Colleges are effectively mini-cities

"I think that you fundamentally misunderstand the markers that make an in-person education higher quality" -> I founded an in-person college with regional accreditation that had a lot more 1:1 and small group teacher time than HYPSM and an average starting salary on par with CS grads from these schools. Our alumni have gone on to become YC founders and can be found at most top tech companies and startups

It is a choice to value exclusivity for exclusivity's sake (eg. withholding JSTOR data from students of colleges who can't afford those costs). The best institutions (eg. YC, Apple) care a lot more about what you can build than what school you got into at age 17

Having personally run a college P&L, this dodges the bigger sunk costs of higher education: 1. Old and expensive to maintain land 2. High cost of living for all staff (weighted heaviest towards faculty) 3. Ancillaries that are revenue negative, _very_ expensive, and inconsequential to the purpose of the education (eg. the lacrosse team and the Polo Club)

It's nice to point fingers at the people who are taking very heavy paycuts to remain in academia, but the result of that finger pointing is devaluing education

The right approach - in my eyes - is to share the land Harvard, Stanford, et. al. sit on with 10x the number of students. This simultaneously increases efficiency of the entire P&L while providing a higher quality of education to everyone

As we've seen with the UC system (and the excellence of IITs + Chinese research universities), high density education can be synonymous with top tier research outcomes - Ivory Towers are not needed

In India (among others), honking is essential to reducing crashes

We often greatly underestimate / undervalue the role of our ears relative to vision. As my film director friend says, 80% of the impact in a movie is in the sound

The biggest surprise I had in attempting to distribute my first Android app is how difficult it is to get beta-testers through the "standard" channels. It requires a 1 week review and 25 beta-users invited by email addresses

In contrast, Apple has a ~48 hour turnaround for reviews before you can upload to TestFlight and distribute a beta with a link

Not sure if I am in some "trusted developer" cohort on iOS but not Android - but the difference was enough for me to stop trying on Android

India and Africa are significantly more optimistic about AI than US and EU

There exists great promise in AI to be an equalizing force, if implemented well

The future is yet to be written

"Swarm of autonomous drones kills 3 buildings of civilians, Silicon Valley is shocked, CEO's offer condolences" is a byline waiting to happen[1]

The administration and the executives will make justifications like: - "We didn't think they would go haywire" - "Fewer people died than with an atomic bomb" - "A junior person gave the order to the drones, we fired them" - "Look at what Russia and China are doing"

Distracting from the fact that the purpose of spending $1.5T/year on AI weapons (technology that has the sole purpose of threatening/killing humans) run by "warfighters" working for the department of war

At no point will any of the decision makers be held to account

The only power we have as technologists seeking "AI alignment" is to stop building more and more powerful weapons. A swarm of autonomous drones (and similar technologies) are not an inevitability, and we must stop acting as if it is. "It's gonna happen anyways, so I might as well get paid" is never the right reason to do things

[1]https://financialpost.com/technology/tech-news/openai-tapped...

This feels like a fair question (perhaps not perfect wording, but no adhominem or disingenuity)

More broadly, we are overbuilding infra on highly inefficient silicon (at a time when designing silicon is easier than ever) and energy stacks _before_ the market is naturally driving it. (with assets that depreciate far faster than railroads). Just as China overbuilt Shenzhen

I have heard (unconfirmed) that the US is importing CNG engines from India for data center buildouts. I loved summers in my youth in Bombay and the parallax background have been great for photography, but the air is no fun to breathe (and does a kicker on life-expectancy to boot)

If we aren't asking these questions here, are they being asked? Don't bite the hand that feeds?

On device models (deepseek-coder, etc) are very good // better than the old way of using stack overflow on the internet. I have been quite productive on long haul flights without internet!

You're an engineer, your goal is to figure stuff out using the best tools in front of you

Humans are resilient, they reliably perform (and throw great parties) in all sorts of chaotic conditions. Perhaps the thing that separates us most from AI is our ability to bring out our best selves when baseline conditions worsen

More entrepreneurs will help with youth unemployment

Welcome to the world by population that speaks English

Developer here. This was a blast to work on over the past few months in collaboration with the Magenta team. Built using a C++/JUCE foundation and a React frontend

Bridging generative music tools and music studios has been a long running theme for the project since 2017: https://www.youtube.com/watch?v=yz-fHidp1M8

Happy to answer any questions :)

We could also just stop breeding genetically modified sentient animals for protein and directly synthesize protein...

My websites have this too with MDX, it's awesome. Reminds me of the old Bret Victor interactive tutorials back around when YC Research was funding HCI experiments

I've been using this for a year, it's really great. Thanks for making it!

uber is like a fine wine. it will appreciate and pay dividends until it bursts when waymo et. al. take over the streets in 20 years

This (particularly the figure 1 illustration) discounts the "distribution" layer for apps

Single app/feature startups will lose (true long before AI). A few will grow large enough to entrench distribution and offer a suite of services, creating defensibility against competitors

The distributors (eg. a SaaS startup that rapidly landed/expanded) will continue to find bleeding edge ways to offer a 6-12mo advantage against foundation models and incumbents

GitLab is a great example of this model. The equivalent bitter lesson of the web is that every cutting edge proprietary technology will eventually be offered free open source. However, there is a commercial advantage to purchasing the bleeding edge features with a strong SLA and customer service

The mistake is to think technology is a business. Business has always been about business. Good technology reduces the cost of sale (CAC) and cost of goods sold (COGS) to create a 85-90% margin. Good technology does not create a moat

Resilient businesses do not rely on singular technology advantages. They invest heavily in long term R&D to stay ahead of EACH wave. Resting on one's laurels after catching a single wave, or sitting out of the competition because there will be bigger waves later, are both surefire ways to lose the competition

There’s been a strong backlash to releasing AI models in my community particularly driven by EA and VC rhetoric that the potential risks far outweigh the benefits

This pushback seems quite surprising to me, since the same claims could be made for the internet, social media, YouTube, etc - the very technologies that put this group into power/wealth

My friends/family in India are far more pro-AI than friends in Silicon Valley / EU. And my artist friends seem to be more excited about it than my VC friends

I think subconsciously, a lot of people currently in power seem to be fearful - perhaps because their power/wealth/way of life will be threatened by a faster rate of change. While those in India seem to see universal translation / accessibility of global knowledge as an equalizing force

These are broad generalizations ofc. But it reminds me of the backlash against nuclear power (despite human life cost of oil/gas being way higher than nuclear) or the CEQA/environmental driven opposition to building more housing in California

I don’t know think individuals are ill intentioned in their focus on AI safety, but they sit in an echo chamber where this technology has more potential to harm them (they have a lot to lose) than benefit them (they are already at the top). So it’s natural to feel threatened and galvanize to slow down the rate of change

I tend to be quite dissatisfied by the current state of the world. Lots of violence (humans/animals), oppression, and gatekeeping of knowledge driven by our colonialist history. So I bias towards faster rate of change that could shift balance of power from western military-industrial complex to global south. AI excitement in global south excites me!

Creators are already curators. Musicians often don’t produce their own sounds, they curate and piece together samples. Designers and software engineers cut and paste from existing work.

The idea of a blank slate creator has been dead long before ML tools were introduced :)

Hey folks! I'm Ashu, the founder of Make School. We're building a new university based in San Francisco and just launched a new online course platform to share curriculum from our CS bachelor's degree for free online.

We'd love to get your feedback on the new course platform as well as the course itself. We're aiming to pair high quality videos with in-browser coding exercises for a combo passive / active learning course.

The platform was built on Gatsby in about a month. Personally, this was my first time writing production code in nearly 5 years, so feeling excited to have launched this :)

This would be a great justification to pause this trading for all clients. The challenge is they didn't limit access for institutional while limiting access for retailers. This discrepancy is core to the feeling that Robinhood is supporting the financial establishment over their community of traders.

Checks are, quite literally the least secure mode of non cash transacting. Yet the banking system is built in a way that checks are often either the fastest way to get someone money, or the only way certain merchants accept payment.

This technology should have been sun-setted years ago, and wire transfers should be free and instant.

Good question! We're definitely leaning stronger on the hands on right now, but are looking to build out more on the math side. Our Dean actually has a PhD in Applied Math rather than CS, so it's in our wheelhouse. It's probably still going to feel closer to Olin than MIT.

You're right that it's impossible to do both at sufficient depth. Ideally we offer students flexibility to push in the direction they want to go. Eventually we'll be working with students who want to go deeper in AI/ML, or potentially other computational STEM fields - they may need stronger math. While others may want to stay closer to product development.

Either way, it's most important to equip students with the mindset and resources to keep learning. Learning should not stop in college, there's no concrete reason why either the math or bench skills can't be picked up later in life - apart from societal bias that tells us to stop learning.