ZeroTalent

o3, reasoner.com, and complex setups of "thinking" workflows for sonnet 3.7, gemini 2.5 pro, and o1-pro

Gemini 2.5 pro is basically free.

Also watsonx, but that's b2b.

LLMs work much better on narrow tasks. They get more lost the more information you introduce. Models are introducing reasoning now which is trying to assert this problem and some models are getting really good at it like o3 or reasoner.com. I have access to both and it looks like, soon, we will have models that become more accurate when we introduce more complexities, which will be a huge breakthrough in AI.

But this is why LLMs are so amazing. They understand context and nuance, and they have reasoning skills now. So you will not get a long list of garbage from a good model.

This is not plagiarism.

These two works express their ideas in very different ways and with very different details, structures, and styles.

That's like saying that Orwell's 1984 is a plagiarism of Yevgeny Zamyatin's We or that Harry Potter is a plagiarism of Ursula K. Le Guin's Wizard of Earthsea.

You are just tackling similar topics. Maybe he took some inspiration from you. You should be proud instead of being butthurt for 15 years.

By the way, the notion that all beings are manifestations of a single consciousness—or that we are all God—is far older than either of you.

That's an interesting take, and it sounds right.

I wonder what the root causes are. Lack of civil education, community strength, and things like these probably play a big part in that division. Educated, clever, happy, balanced people would not let someone like Trump (grab em by the pussy, etc) come even an inch to the any governing body.

There is a better way that I'm using:

1. Cursor Pro with Sonnet to implement things the Cursor way.

2. Install the Gemini Code extension in Cursor.

3. Install the Gemini Coder Connector Chrome extension: https://chromewebstore.google.com/detail/gemini-coder-connec...

4. Get the free aistudio.google.com Gemini API and connect the extensions.

5. Feed your codebase or select files via the Cursor extension and get the implementation from aistudio.google.com.

I prefer having Sonnet implement it via Cursor rather than Gemini because it can automatically go through all the linting/testing loops without my extra input, run the server, and check if there are no errors.

I believe their point was that they are writing about what people want to read (a new AI breakthrough), possibly embellishing or cherry-picking results, although we can't prove/disprove it easily.

This approach yields more upvotes and views on their website, which ultimately leads to increased conversions for their tool.

There is no societal solution for this right now, as far as I understand.

I'm only saying that places like these exist—smaller, walkable, pleasant, safe cities designed for families and communities. These are just a few that I have found in my life.

And, keeping it real, it's reserved for people with money—if there would be good jobs there, too many people would come, and the large population problems would emerge.

It is sad.

Writing this reminds me of the short story by Ursula K. Le Guin—The Ones Who Walk Away from Omelas.

Do you not consider people who voted for a proto-fascist president, if not evil, at least responsible for his wrongdoings? Do you really think we can take the middle road here? Like the two-state solution with Israel?

I don't.

This is the paradox of tolerance. The fascists will eat everyone alive if they are allowed to freely express their "opinions" and implement them on a societal scale.

I have left my country because of this.

And then the government wants us to play by its rules. Sure /s

Apple is pretty awesome at privacy compared to everything else.

> As if bots weren't enough of a problem, imagine when social media is just people clicking on buttons: "write a funny response", "write a comment in disagreement", "write 'same'", etc.

There are already extensions for this purpose.

replai.so and dozens others

I hear you, but some people just want an easy high-paying job where they essentially work a few hours a week. Not everyone wants to fold proteins.

It's that we open Hacker News right after we wake up and post comments before our brains are in full power, hence the mistakes. Coffee is not a necessity, but most people drink it. I believe it's not exactly about caffeine but letting the brain 30 minutes or so to "wake up" to stop making mistakes. At least, that's how it works in my case.

Simple example: My company is using Gemini to analyze every 13F filing and find correlations between all S&P500 companies the minute new earnings are released. We profited millions off of this in the last six months or so. Replicating this work alone without AI would require hiring dozens of people. How can I not be bullish on LLMs? This is only one of many things we are doing with it.

I do not understand how you can be bearish on LLMs. Data analysis, data entry, agents controlling browsers, browsing the web, doing marketing, doing much of customer support, writing BS React code for a promo that will be obsolete in 3 months anyway.

The possibilities are endless, and almost every week, there is a new breakthrough.

That being said, OpenAI has no moat, and there definitely is a bubble. I'm not bullish on AI stocks. I'm bullish on the tech.

they have access to o3, I do. Thousands of people do(tens of thousands at this point?). Come on. Compare to SOTA, when you're saying it's the best AI you have.

This makes a lot of sense. Some people will think twice about writing an article like that again. (unfortunately)

I've seen it only because I use hn.algolia.com to read hacker news.

This one is flagged. Proves the point.

Legally? :D

Also, why not compare to GPT-o3 in the benchmarks?

Kinda similar, but when I looked at the finances, I was surprised by how much money we're getting from places like the Cayman Islands, Switzerland, and the Emirates.

Same with AWS SES, in my experience.

Also we barely understand how cognition works, AFAIK.

it's tough to fight against fascism without the use of violence or a very heavy threat of violence. I think that most major changes in history show this to be true.

barely anyone would even mention a regular protest in front of dealerships.

Just neofascism.

I don't think true FSD is possible without lidar.

You can derive the "true price" with a certain probability. Of course, unexpected things can happen, like Musk being assassinated, but the stock market is not completely unpredictable.

And the thing is, I don't have to be right about Tesla because I'm not making a single bet on that company. I'm making hundreds of bets using my methodology, and it's enough that I'm right 50-something% of the time (or even less than 50%, if we're talking Options trading).

That's what the whole stock market game is about. Hence, some investment companies, like Berkshire Hathaway, consistently make more than others over the long term.

I'm claiming this based on fundamental analysis of Tesla's sales/net profits/ROIIC and other metrics, which I take from their 13F fillings (AKA quarterly earning reports) compared to other car manufacturers.

Some stocks are overpriced, and others are undervalued. The inclusion in the S&P500 alone is just 1 of the factors.

In my opinion, Apple and NVIDIA are significantly undervalued based on their fundamentals, even though they make up a gigantic % of the SPX.