Art9681

5 ms·

61.
by Art9681 5mo ago
Except the latency is significant and not suitable for clients with advanced agent features. The experience between using a frontier model via first party API and the best open weight models via OpenRouter is night and day. Can't get a
62.
by Art9681 5mo ago
Americans don't care what the rest of the world says. It's a privilege that comes from a position of power. In any case, the Iranian people themselves have protested their own regime many times, so no one needs to listen to the US
63.
by Art9681 5mo ago
The US is winning the AI race in all matters.
64.
by Art9681 5mo ago
It wouldn't have the same appeal if they reported seeing "larger than average size squids we didnt know existed". Every story was embellished in a world without pocket cameras. And the further you go back, the more grand the
65.
by Art9681 5mo ago
This is the truth. If these "phenomena" where real we wouldn't question it. None of these reports would be necessary. It would just be common knowledge. It's the modern day equivalent of Big Foot or Nessie and its releva
66.
by Art9681 5mo ago
Behold. The most comprehensive reports of high tech equipment malfunction in the history of humanity. It is no surprise the most incompetent administration in the history of the United States would release this obvious PsyOp to the public i
67.
by Art9681 5mo ago
Any rando can publish research nowadays. It means nothing. Just like "X country published N research papers last year". It is noise. In a world where it was required to attach age, experience level, and country of origin to every
68.
by Art9681 5mo ago
Remind yourselves that most research papers are written by career students with no real world practical experience. That is all.
69.
by Art9681 6mo ago
Counterpoint: If humans shipped perfect products they would no longer havejobs. The majority of time spent in an organization is fixing problems humans caused. For good reasons and bad excuses. We are not machines. What we, collectively as
70.
by Art9681 6mo ago
We can because the reality is that America has led in AI since the beginning and has had the best frontier models. It's not like some other country held the top spot for any given period of time. No one in Europe or China. I'd giv
71.
by Art9681 6mo ago
A junior tinkering in their garage in domains they have little experience executed a flawed test and decided to call it a benchmark. It's extremely common nowadays because words dont mean anything anymore. The forums that used to be fi
72.
by Art9681 6mo ago
I have no dog in this fight but a friendly reminder that temperature and weather are not synonymous.
73.
by Art9681 7mo ago
Aside from the fabricated drama and the trend chasing, OpenAI still has the best overall model and API service. Anthropic is really good, no doubt. But gpt-5.4 is a better model than even Opus, even if its a marginal advantage. I use both.
74.
by Art9681 7mo ago
That is just idealism. Being "open" doesnt get you any advantage in the real world. You're not going to meaningfully compete in the new economy using "lesser" models. The economy does not care about principles or et
75.
by Art9681 7mo ago
This was an experiment conducted during gpt-3.5 era, and again during the gpt-4 era. There is a reason it is not a common/popular technique.
76.
by Art9681 7mo ago
How convenient of them to compare themselves to the last generation Opus and GPT models to make their model look better than it really is.
77.
by Art9681 7mo ago
Can't we simple parse and remove any style="display: none;", aria-hidden="true", and tabindex="1" attributes before the text is processed and get around this trick? What am I missing?
78.
by Art9681 7mo ago
That solution can be recreated by a skilled AI boosted senior platform engineer in a few days and parity achieved in a few weeks. Nothing of value was lost.
79.
by Art9681 7mo ago
Is "how programmers work" a useful and provable metric? No? Then it belongs in philosophy discussions. How you work and how I work is different. Your work may have ended up in the LLM training and my work did not. Or vice versa. C
80.
by Art9681 7mo ago
I enjoy the journey too. The journey is building systems, not coding. Coding was always the most tedious and least interesting part of it. Thinking about the system, thinking about its implementation details, iterating and making it better
81.
by Art9681 8mo ago
Should've used Go.
82.
by Art9681 8mo ago
It's going to be faster no matter what. My M3 MAX prints tokens faster than I can read for the new MoE models. It's the prompt processing that kills it when the context grows beyond a threshold which is easy to do in the modern ag
83.
by Art9681 8mo ago
128GB RAM maximum? What a shame. I was ready to spend 8k on a laptop. My M3 MAX 128GB will suffice for now.
84.
by Art9681 8mo ago
Of course the US is going to do this and of course its in Anthropics best interest to comply. Right now China is flooding HuggingFace with models that will inevitably have this capability. Right now there are hundreds of models being hosted
85.
by Art9681 8mo ago
I'm going to claim that the majority of those users are optimizing for cost and not correctness and therefore the quality of data collected from those sessions is questionable. If you're working on something of consequence, you&#x
86.
by Art9681 9mo ago
Their models are specifically trained for their tools. For example the `apply_patch` tool. You would think it's just another file editing tool, but its unique diff format is trained into their models. It also works better than the gene
87.
by Art9681 9mo ago
A reminder the insatiable appetite of the American consumer is a big part of what props up a lot of other economies.
88.
by Art9681 9mo ago
The world wasn't globalized then. It's quite literally, an entire new world today. The collapse of the US will touch every corner of the world.
89.
by Art9681 9mo ago
What makes you think the next generation models won't be explicitly trained to prevent this, or any other pitfall or best practice as the low hanging fruit fall one by one?
90.
by Art9681 9mo ago
The author must have taken inspiration from Orbstack which is itself a MacOS app for managing containers, VMs and a local Kube cluster. The first image in the README shows they've selected the cluster Orbstack provisions as the one bei

More ›