HNHacker News
TopNewBestAskShowJobs

geysersam

2,843 karma · joined August 21, 2021

submissionscomments
geysersam··on The Socialist Temptation of Sam Altman
Why do you say that?
geysersam··on Matrix Orthogonalization Improves Memory in Recurrent Models
They said "there are algorithmic changes that remains to be discovered" and you said they bought into the idea that "all we need is more context". Seems like opposites to me.
geysersam··on Statistics that live in your SQL
Looks great!

One minor correction - the `summarize` function in duckdb can also be used in CTEs etc.

But you have to wrap the `summarize` in a `from` clause like this:

  with
    some_table as (from range(10)),
    x as (from (summarize some_table))
  from x;
geysersam··on Curl will not accept vulnerability reports during July 2026
This is the HTTP/1.1 standard: https://datatracker.ietf.org/doc/html/rfc2616

Then there are also HTTP/2 and HTTP/3.

That's just HTTP, curl supports 27 other protocols.

geysersam··on EU Commission looking at practical consequences of Anthropic decision
Call it progress, call it "economic leverage", a very inflated stock market is no good measure for either.

It's true the US economy has grown faster than the European in the last two decades, for sure, I agree with you. It's just that the apparent magnitude of the difference depends on what metric you choose, and I just mean that the value of the stock market seems like a poor metric.

geysersam··on EU Commission looking at practical consequences of Anthropic decision
To be fair the value of the stock market is not a good indicator of progress. The US stock market is incredibly inflated.

Besides, it's a bit strange to argue that it's impossible to make a change, and as proof of that take the fact that there's been a big change over the last 20 years.

geysersam··on How to earn a billion dollars
To add to that: There are also "compounding effects" from investments made by tax money, just as for any other investments, the difference is that the compounding gains are collectively owned and not controlled by an individual.
geysersam··on How to earn a billion dollars
That's a misunderstanding of the original argument. It's not about billionairs "doing something bad". The claim is that the only way to become a billionaire is to benefit from an unfair economic system.
geysersam··on How to Earn a Billion Dollars
Well put!
geysersam··on How to earn a billion dollars
The word "extracted" does not betray a belief that value cannot be created. You can "extract" value that is created just as you can extract value that was there already. The question is not whether or not value was created, the question is who deserves to control the value that was created.

The fact is the billionaire managed to extract value from the market. The ethical question is: who deserves to get the value that was created by the market? The answer could be "the founder" but it could also be the funder, the worker, the customer, the political structure that enables the market economy, the mother of the funder who raised them to be hard working, the nurse that treated the founders minor illness in an early stage and prevented it from causing a physical disability, etc.

geysersam··on Claude Fable is relentlessly proactive
This is the worst thing about current AI agents. They never ask questions. The prompt has to be pixel perfect and unambiguous or they'll happily run away doing something ridiculous.
geysersam··on Chrome is looking to permanently drop MV2 extension
Hah

> 6. You can make money without doing evil

implies that they're doing it for fun then I guess?

geysersam··on Chrome is looking to permanently drop MV2 extension
Finally Firefox will get a 30% usage share!
geysersam··on Job: Head of Stonehenge
By your definition 95% of the world population live in 'poor' countries. I guess if that's how you want to use the word that's up to you, but people outside of your bubble will literally not understand what you are saying.
geysersam··on Job: Head of Stonehenge
Let's not be delusional. The UK is not a poor country, and 64K is low by US tech standards but it's good by any other measure.
geysersam··on Zig Structs of Arrays (2024)
Genuine question: why do you think game engines need operator overloading? I mean, what's wrong with generic functions like add, multiply, dot etc.
geysersam··on The Pentagon is running an AI propaganda mill targeting Latin America
The context of this discussion is Brazil and the intentions of the socialist Brazilian government. Are you claiming that the socialist party in Brazil intends to dismantle capitalism? In case you are unaware, countries can have capitalist economic systems and still have significant socialist traits like: large public sectors, state enterprises for natural monopolies and important industries, subsidized education and healthcare, etc.
geysersam··on The Pentagon is running an AI propaganda mill targeting Latin America
There are lots of socialist countries in the world that are doing fine.
geysersam··on I'm skeptical about efforts to revolutionize schooling
I think kids have a right to be exposed to these ideas. Stratifying society even further, only exposing select sections of the population to advanced ideas in philosophy or science, will not help build a solid democratic society.
geysersam··on Uber's $1,500/month AI limit is a useful signal for AI tool pricing
Current AI datacenter/model development investment rate is roughly 1T/year. That's a lot. But the US economy is 33T/year. So the investment pays back (roughly) over ten years if, each year, the AI investments increase overall productivity by 0.6%, assuming the AI companies can capture half of the value of that productivity gain.

> „[AI vendors are] paying for a fixed cost with a depreciating commodity“

That's just a confusing way to say you don't think future models will be worth the development costs. Because if future models are significantly better, why would the price of tokens to access those models deprecate?

geysersam··on 1-Bit Bonsai Image 4B Image Generation for Local Devices
There's always existed misinformation in text and in images. It's been possible to manipulate photographs for as long as photography has been around. It's becoming easier for sure, but it's not really a qualitative change. Trusting anything you saw on the internet twenty years ago would have been as ridiculous as it is today.
geysersam··on Cloudflare Turnstile requiring fingerprintable WebGL
8 billion checks per day sounds on the low end. I can imagine it being ten or hundred times more. That still seem pretty fine though. On the other hand, it's hard to see that such a modest energy cost would dissuade any attacks.
geysersam··on MCP is dead?
Interesting, thanks!
geysersam··on MCP is dead?
All agents have some basic tool capabilities, I meant using tools such as curl or python etc to access an api, the same way a human would.
geysersam··on MCP is dead?
Seems to me that you're saying the MCP is a simplified API with good documentation geared towards agents. But if that's the case, could you not exposes the simplified interface as part of the API, instead of exposing it in MCP?
geysersam··on MCP is dead?
But what is the advantage of MCP compared to having the agents access the API directly?
geysersam··on MCP is dead?
What advantage did you find in MCP vs a traditional API?
geysersam··on Quack: The DuckDB Client-Server Protocol
I don't think that's correct. DuckDB already supports concurrent writes within one process. I don't see why this would suddenly serialize all writes.
geysersam··on Quack: The DuckDB Client-Server Protocol
Agreed, that does seem a bit underwhelming. Hopefully there are some performance gains to be made before the production release in september.
geysersam··on Quack: The DuckDB Client-Server Protocol
They mention in the benchmarks section that the network they're on is a "up to" 15 Gbps connection. So to max out 50GB/s is not realistic.

I agree they should have also listed the compressed size of the table instead of only mentioning the CSV size. But the compressed dataset is probably not smaller than 1/10 of the CSV size. If that's the case they're transferring ~8GB in 4.6 s on a 2GB/s (15Gbps) connection. Seems pretty close to max.

← PreviousPage 3 of 34Next →