benchmarkist

12 ms·

91.
by benchmarkist 2y ago
Interpretability research is basically a projection of the original function implemented by the neural network onto a sub-space of "explanatory" functions that people consider to be more understandable. You're right that the
92.
by benchmarkist 2y ago
Eventually the AI folks will discover type systems. The graph system outlined in the article is a very rudimentary one but a step in the right direction.
93.
by benchmarkist 2y ago
> The advocates of what was then called "hard A.I.," to the dismay of their opponents, were painting a triumphalist picture of the future of computer intelligence. As the discussion went on, all semblance of logical argument wa
94.
by benchmarkist 2y ago
Title is misleading. It's not a vaccine. They genetically engineer the parasite to reduce its lifetime so that it does not persist in the liver. It's not a long-term solution because the parasite can potentially evolve and overcom
95.
by benchmarkist 2y ago
I recommend you list them instead of telling random strangers about the helpfulness of their responses.
96.
by benchmarkist 2y ago
Who is footing the bill for the bankruptcy?
97.
by benchmarkist 2y ago
Losses are socialized, profits and bonuses are privatized. That's how the current system is set up to work.
98.
by benchmarkist 2y ago
Everyone is entitled to their opinions.
99.
by benchmarkist 2y ago
That's a possible interpretation but the verbosity of the original statement hinders understanding the actual intent which is I why I recommended simplifying and removing unnecessary redundancies.
100.
by benchmarkist 2y ago
In psychology that's called projection. I highly recommend looking into it.
101.
by benchmarkist 2y ago
I would recommend you take your own advice.
102.
by benchmarkist 2y ago
https://dlicata.wescreates.wesleyan.edu/pubs/amlh14patch/aml...
103.
by benchmarkist 2y ago
As long as it is militarily and commercially viable then the number of satellites will continue increasing, regardless of what academics have to say about collision rates. As per usual this is a coordination problem and in case people have
104.
by benchmarkist 2y ago
Every quantum algorithm is a unitary operation in a Hilbert space. If you want to understand the theory then you will have to do the actual work of learning about Hilbert spaces and unitary operators.
105.
by benchmarkist 2y ago
Never expect factual accuracy from large language models. No one actually knows how to make neural networks which always conform to a well-defined knowledge graph and so that means you should always assume all of the output contains subtle
106.
by benchmarkist 2y ago
Networking over websockets with tailscale is a nice touch.
107.
Llama 3.1 405B now runs at 969 tokens/s on Cerebras Inference(cerebras.ai)
427 points by benchmarkist 2y ago|156 comments
108.
by benchmarkist 2y ago
What are the technicalities?
109.
by benchmarkist 2y ago
Money is fake so we can print as much of it as we want. The problem is that innovation can't be bought with money. Newton did not invent calculus because he wanted to get rich, he invented calculus to understand the universe. Money is
110.
by benchmarkist 2y ago
I guess that's why everyone is in a rush to develop AI, artificial wombs, genetic engineering, and robots but given the scale of the ecological damage I'm not sure what exactly the survivors are going to do with the entire mess.
111.
by benchmarkist 2y ago
It's based on my knowledge of mathematics and software engineering. I have a graduate degree in math and I have written code for more than a decade in different startups across different domains ranging from solar power plants to email
112.
by benchmarkist 2y ago
The history of life is a history of extinctions and I don't think humanity is an exception to that rule.
113.
by benchmarkist 2y ago
Technology is not going to get us out of this mess.
114.
by benchmarkist 2y ago
Very likely.
115.
by benchmarkist 2y ago
The linked article is purely for marketing purposes and contains nothing of technical value or substance.
116.
Modeling urban growth shows that cities develop in ways similar to tumors(phys.org)
1 points by benchmarkist 2y ago|0 comments
117.
by benchmarkist 2y ago
This is a great article but the main principle at YC is to assume that technology will continue progressing at an exponential rate and then thinking about what it would enable. Their proposals are always assuming the startups will ride some
118.
by benchmarkist 2y ago
Test-time compute is also known as beam search but I guess it's good to see the neural network folks keeping the tradition of rediscovering old techniques and giving them different names.
119.
by benchmarkist 2y ago
Those are all pretty good. I'm a fan of Futurama so branniganisms/brannigans is a good label in my opinion.
120.
by benchmarkist 2y ago
Hallucinatory advice is inherent to all LLMs. Nothing that comes out of an LLM can or should be trusted unless you can verify the output some other way via a trusted 3rd party, preferably human one.

More ›