FYI, 3.8 Flash is available on aistudio.google.com (along with all of their other models)
But yeah, they really dgaf about gemini.google.com -- I dropped that sub in April when it was clear OAI and Anthropic had lapped them
2,927 karma · joined August 1, 2016
But yeah, they really dgaf about gemini.google.com -- I dropped that sub in April when it was clear OAI and Anthropic had lapped them
The Theory of Constraints - AI Era
[0] https://en.wikipedia.org/wiki/Theory_of_constraints
[1] https://www.goodreads.com/book/show/113934.The_Goal
[2] https://www.goodreads.com/en/book/show/17255186-the-phoenix-...
Zach is undoubtedly smart but for anyone who is not an SV insider, they would listen to that podcast they same way you are looking at this comment and wonder if it's all one big joke.
I was also never quite sure why he was in the orbit of influence and power. He had a failed bitcoin mining hardware company and that's about all I know of, and yet he kept seeming to fail upwards.
Oh and don’t forget that error message being returned when you try to call the API is because you didn’t give your project the proper permissions in google cloud console. What permissions do you need? ¯\_(ツ)_/¯
Google Cloud Console feels like being stuck in the seventh circle of hell.
This isn't really unknown either. There's a very good story anyone can look up about Dr. V in India and what it took for him to actually get the eye care he wanted to provide to the people who needed it.
In the digital world many of us know you want to deeply understand your user and design with them in mind. Same thing here in the meat space.
> The competition is therefore all the other activities that compete for the rapidly growing “discretionary time” of a population
His examples were bowling ball manufacturers competing with lawn care companies, but the idea is the same, go up an abstraction layer, and the competition is for time.
My perspective from someone who wants to understand this new AI landscape in good faith. The water issue isn't the show stopper it's presented as. It's an externality like you discuss.
And in comparison to other water usage, data centers don't match the doomsday narrative presented. I know when I see it now, I mentally discount or stop reading.
Electricity though seems to be real, at least for the area I'm in. I spent some time with ChatGPT last weekend working to model an apples:apples comparison and my area has seen a +48% increase in electric prices from 2023-2025. I modeled a typical 1,000kWh/month usage to see what that looked like in dollar terms and it's an extra $30-40/month.
Is it data centers? Partly yes, straight from the utility co's mouth: "sharply higher demand projections—driven largely by anticipated data center growth"
With FAANG money, that's immaterial. But for those who aren't, that's just one more thing that costs more today than it did yesterday.
Coming full circle, for me being concerned with AI's actual impact on the world, engaging with the facts and understanding them within competing narratives is helpful.
It’s very rare to find a local coffee shop in the U.S. with a drive-thru.
If I'm running a lawn care company in the desert I can get all those annoying details right and still be unsuccessful. So strategy is not opening a lawn care company in the desert.
If you think I'm missing something you are saying, please let me know!
If you're in the wrong market or building for the wrong customers you can execute brilliantly on everything you mentioned and it won't matter. The only thing that matters is product market fit or finding it if you don't have it. That's what I see as unsaid in your parent's comment about "execute what though?"
I don't have a test like that on hand so I'm really curious what all you prompted the model, what it suggested, and how much your knowledge as a SWE enabled that workflow.
I'd like a more concrete understanding if the mind blowing nature is attainable for any average SWE, an average Joe that tinkers, or only a top decile engineer.
1. thefacebook
2. transition from desktop to mobile
3. building the machine that facebook became
4. buying out or building feature parity with competitors that took FB from its IPO market share of $104 billion to today's market cap of $1.89 trillion.
Has he innovated successfully since the o.g. thefacebook? Not really. Metaverse fell flat on its face. Hardware efforts over two decades have gained no meaningful traction. AI is a mess.
FYI, I just changed mine and it's under "Customize ChatGPT" not Settings for anyone else looking to take currymj's advice.
A story with intrigue that chronicles the why and how Microsoft ended up extracting the most value from the PC revolution instead of the hardware makers and of course, why that was DOS instead of CP/M.
I liked the oral history nature of this podcast, walking me through things that preceded me in technology, and then things that I lived through like the 90's internet.
https://www.internethistorypodcast.com/2016/03/the-man-who-c...
In practice when I have seen ARC brought up, it has more nuance than any of the other benchmarks.
Unlike, Humanity's Last Exam, which is the most egregious example I have seen in naming and when it is referenced in terms of an LLMs capability.
OpenAI and Microsoft are at a standoff over the terms of the startup’s $3 billion acquisition of the coding startup Windsurf, the people said. Microsoft currently has access to all of OpenAI’s IP, according to their agreement. It offers its own AI coding product, GitHub Copilot, that competes with OpenAI. OpenAI doesn’t want Microsoft to have access to Windsurf’s intellectual property.
> tl;dr on Section 174, Research & Experimentation costs went from being fully deductible in the year incurred to being deductible over a 5 year period.
Larger tax bills and a tightening on what roles/activities are deductible as R&E are likely what OP is pointing at with his comment.
To the best of my non-inside baseball research, Section 174 changes were simply one part of a package of revenue generating measures to offset the large tax cuts from the broader tax act they were a part of.
The changes came from The Tax Cuts & Jobs Act of 2017 that was introduced to the House of Representatives by Congressman Kevin Brady (R) Texas. The bill passed both houses of Congress along party lines. Then President Trump signed the bill into law. Section 174 changes did not take effect until 2021.
https://news.ycombinator.com/threads?id=heymijo&next=4332098...
I don’t know Varun (their founder/CEO) personally but I get highly competent vibes from him. I’d let my skeptical self lean on your optimistic take.
A high-functioning team is going to have at least one person who does this. For a perpetually high functioning team this is going to be second nature.
That is curious. Things are moving so quickly right now. I typed out a few speculative sentences then went ahead and asked an LLM.
Looks like Cerebras is responding to the market and pivoting towards a perceived strength of their product combined with the growth in inference, especially with the advent of reasoning models.
Why?
For each new word a transformer generates it has to move the entire set of model weights from memory to compute units. For a 70 billion parameter model with 16-bit weights that requires moving approximately 140 gigabytes of data to generate just a single word.
GPUs have off-chip memory. That means a GPU has to push data across a chip - memory bridge for every single word it creates. This architectural choice, is an advantage for graphics processing where large amounts of data needs to be stored but not necessarily accessed as rapidly for every single computation. It's a liability in inference where quick and frequent data access is critical.
Listening to Andrew Feldman of Cerebras [0] is what helped me grok the differences. Caveat, he is a founder/CEO of a company that sells hardware for AI inference, so the guy is talking his book.
[0] https://www.youtube.com/watch?v=MW9vwF7TUI8&list=PLnJFlI3aIN...
There's a sweet spot for an app that is inaccurate with a market that wants it but doesn't understand how inaccurate it is.
Kind of like how I could vibe code an app, get it to "work", think it's great and be ignorant of the many ways it will break or isn't working that a knowledgeable developer could.