behnamoh

41 ms·

931.
by behnamoh 2y ago
This is Apple Intelligence the way it was supposed to be ("AI for the rest of us"), but Apple just doesn't "get" AI, so here we are—the only platform provider that is taking the correct approach to AI is Google.
932.
by behnamoh 2y ago
Is it a replacement for Jinja2 templates? I use them a lot in LLM pipelines (e.g., to fill in the system prompt and provide more context).
933.
by behnamoh 2y ago
The naming of these chips (GPUs, CPUs) is kinda badass: Ironwood, Blackwell, ThreadRipper, Epyc, etc.
934.
by behnamoh 2y ago
It seems companies figured introducing "protocols" or standards helps their business because if it catches on, it creates a "moat" for them: imagine if A2A became the de facto standard for agent communication. Since Goog
935.
by behnamoh 2y ago
It seems the industry as a whole just forgot about prompt injection attacks because RLHF made models really good at rejecting malicious requests. Still, I wonder if there have been any documented cases of prompt attacks.
936.
by behnamoh 2y ago
Not having cons lists makes Janet non-lispy?
937.
by behnamoh 2y ago
in that case Python is also a Lisp because it has: car, *cdr = [1, 2, 3, 4]
938.
VB, Vb.net, VBA, VBScript: Which Visual Basic Is Which?(mobilize.net)
2 points by behnamoh 2y ago|0 comments
939.
by behnamoh 2y ago
I use this: I don't want to drastically change my current code, nor do I like being told to create several new files and numerous functions/classes to solve this problem. I want you to think clearly and be focused on the tas
940.
by behnamoh 2y ago
What I don't like about LLMs is that people keep re-inventing the wheel over and over. For example, we've been able to control browsers using GPT for about 2 years now: - https://github.com/mayt/BrowserGPT -
941.
by behnamoh 2y ago
3.7 is like a wild horse. you really must ground it with clear instructions. it sucks that it doesn't automatically know that but it's tameable.
942.
by behnamoh 2y ago
Continue and several other extensions do what Cursor does and they're extensions. No excuses from Cursor are acceptable.
943.
by behnamoh 2y ago
tbh I'm glad about this, even if it breaks Cursor and other unnecessary forks of VSC. there's a difference between forking something to improve it and give back to the community under the same license vs what Cursor is doing.
944.
Festivitas for macOS(mastodon.social)
1 points by behnamoh 2y ago|0 comments
945.
by behnamoh 2y ago
I have Apple Silicon and it's the worst when it comes to prompt processing time. So unless you want to have small contexts, it's not fast enough to let you do any real work with it. Apple should've invested more in bandwidth,
946.
by behnamoh 2y ago
does the entire model have to be loaded in VRAM? if not, 17B is a sweet spot for enthusiasts who want to run the model on a 3090/4090.
947.
by behnamoh 2y ago
I wonder why they chose 4090 instead of 5090.
948.
Ask HN: Is 5090 a viable option for building GPU clusters?
1 points by behnamoh 2y ago|1 comments
949.
Mac: It Just Works–Terribly(medium.com)
3 points by behnamoh 2y ago|0 comments
950.
by behnamoh 2y ago
monads are the MCP of functional programming—no one really knows what they are but everyone writes an article about them using analogies that break when you actually use them in practice.
951.
by behnamoh 2y ago
Anthropic doing more open AI work than the Ghibli-AI-images company.
952.
by behnamoh 2y ago
I'm glad this exists. It ruins the day for Trelis who took the open-source and free Llama and made it commercial by giving it function calling abilities: https://huggingface.co/Trelis/Meta-Llama-3-70B-Instruct-func
953.
by behnamoh 2y ago
It's just that I don't any example of an adult genius who used to be a prodigy. The most obvious counter example was Einstein.
954.
by behnamoh 2y ago
easy: how many genius people do you know who were also prodigies? early intelligence only gets you so far, the rest depends on hard work, passion, etc.
955.
by behnamoh 2y ago
prodigies are amazing, but I often wonder what they end up doing later in life when the intelligence gap between them and their peers converges to zero.
956.
by behnamoh 2y ago
Google's announcements are mostly vaporware anyway. Btw, where is Gemini Ultra 1 ? how about Gemini Ultra 2?
957.
by behnamoh 2y ago
good grief! people are okay with it when OpenAI and Google do it, but as soon as open source providers do it, people get defensive about it...
958.
by behnamoh 2y ago
so it's hot code reloading on steroids? really neat idea, I'm working on a similar Lisp-style language and will probably adopt this idea (and cite you of course). In my language, you can redefine any symbol, even numbers, so (def
959.
by behnamoh 2y ago
Surprisingly, Gemini (Thinking) doesn't do that—it thinks very formally, as if it's already formed its response.
960.
Apple Silicon performance hasn't seen much improvement over the years(twitter.com)
6 points by behnamoh 2y ago|11 comments

More ›