thimabi

6 ms·

61.
by thimabi 1y ago
What I’d most like in future MacBooks is the continuation of increases in memory size and bandwidth. Apple has carved out a niche for itself in the local LLM space, yet it continues to overcharge for RAM and under-deliver in terms of bandwi
62.
by thimabi 1y ago
No, they finally included Codex usage in the subscription pricing.
63.
by thimabi 1y ago
Poor OpenAI workers, they worked so hard for the GPT-5 release and now discussions about the model are side by side with discussions about their badly-done graphs. I don’t believe they intentionally fucked up the graphs, but it is nonethele
64.
by thimabi 1y ago
> Did anyone yet hit the usage limits to report back how many messages are possible? 10 messages every 5 hours on GPT-5 for free users, then it uses GPT-5-mini. 80 messages every 3 hours on GPT-5 for Plus users, then it uses GPT-5-mini (
65.
by thimabi 1y ago
I personally hated this decision. Of course, I know that having a line-up of tons of models is quite confusing. Yet I also believe users on the paid plan deserve more options. As a paying user, I liked the ability to set which models to use
66.
by thimabi 1y ago
Oh, no, I do understand the value of long-form and deep human communication. I’m an avid book reader, for instance, and I actually prefer longer narratives. What I don’t agree with is being needlessly verbose in circumstances in which the o
67.
by thimabi 1y ago
One of the things that makes me hopeful for the future of LLMs is precisely this: humans are needlessly verbose, and LLMs can cut through the crap. I expect smaller models to become incrementally better at compressing what truly matters in
68.
by thimabi 1y ago
> Large models are for querying the model > Small models are for querying the context I respectfully disagree. My experience is that large models are capable of understanding large contexts much better. Of course they are more expensi
69.
by thimabi 1y ago
That is so cool! Thanks for the Gemini team for working on that, a great and innovative feature. Just a heads up: as I tried to print several stories to PDF, most times one of the generated images did not appear on the PDF. It’s surely a bu
70.
by thimabi 1y ago
I’m not throwing the towel on Ollama yet. They do need dollars to operate, but still provide excellent software for running models locally and without paying them a dime.
71.
by thimabi 1y ago
My guess is that’s the lowest price point that provides a modicum of profitability — LLMs are quite expensive to run, and even more so for providers like Ollama, which are entering the market and don’t have idle capacity.
72.
by thimabi 1y ago
I think it’s both of them, as well as an attempt to compete with other makers of open-weight models. OpenAI certainly isn’t happy about the success of Google, Facebook, Alibaba, DeepSeek…
73.
by thimabi 1y ago
I mean, that’s great, but it would be even better if MacOS required explicit permissions before allowing an app to place an icon in the menu bar.
74.
by thimabi 1y ago
It works fine, although with a bit more latency than non-local models. However, swap usage goes way beyond what I’m comfortable with, so I’ll continue to use smaller models for the foreseeable future. Hopefully other quantizations of these
75.
by thimabi 1y ago
It could be, but there’s so much hype surrounding the GPT-5 release that I’m not sure whether their internal models will live up to it. For GPT-5 to dwarf these just-released models in importance, it would have to be a huge step forward, an
76.
by thimabi 1y ago
Open weight models from OpenAI with performance comparable to that of o3 and o4-mini in benchmarks… well, I certainly wasn’t expecting that. What’s the catch?
77.
Intel struggles with key manufacturing process for next PC chip, sources say(reuters.com)
6 points by thimabi 1y ago|1 comments
78.
by thimabi 1y ago
I’ll be waiting expectantly for the post mortem of this. How ironic would it be if this issue was caused by a pull request itself?
79.
by thimabi 1y ago
Of course! It just saddens me to know that users who are incessantly bombarded with ads, or who fall prey to invasive tracking, or who believe in AI hallucinations… probably won’t have access to a better search experience.
80.
by thimabi 1y ago
There are certain things that Kagi gets very right. Having the ability to (de)prioritize websites, or to right-click to save images, or to automatically rewrite website URLs… Still, I think paying for search remains a taboo and this is unli
81.
I dumped Google for Kagi(arstechnica.com)
362 points by thimabi 1y ago|242 comments
82.
by thimabi 1y ago
That’s in fact one of the gripes I have with certain MacOS software. It would be far better if menu bar icons were opt-in rather than opt-out. The average non-technical user eventually ends up having tons of these icons in the menu bar.
83.
by thimabi 1y ago
> Why are these companies not upfront about what the limits are? Most likely because they reserve the right to dynamically alter the limits in response to market demands or infrastructure changes. See, for instance, the Ghibli craze that
84.
by thimabi 1y ago
> I find it astonishing that the same company providing free usage of their top models to everybody via AI Studio is nickel-and-diming their actual customers like that. I agree that’s not a good posture, but it is entirely unsurprising.
85.
China summons Nvidia over backdoor security concerns with AI chips(washingtonpost.com)
3 points by thimabi 1y ago|0 comments
86.
by thimabi 1y ago
I’ve been facing a similar issue with my MacBook Pro with Apple Silicon. While sleeping with an SSD connected, it seems to wake up periodically and activate the drive to do something. The result? Both the laptop and the SSD eventually overh
87.
by thimabi 1y ago
It came very late indeed! By now, I’m already used to LM Studio as a UI for local LLMs… it even seems to have more features than ollama. But I liked to know that ollama developed a GUI as well — more options is always better, and maybe it w
88.
by thimabi 1y ago
Integration with browsers and websites, because these password managers are based on autofill rather than copying and pasting passwords.
89.
by thimabi 1y ago
For sure! I haven’t seen any other big AI provider with features and UIs as polished as the OpenAI ones. I believed competitors would rush to copy all great things that ChatGPT offers as a product, but surprisingly that hasn’t been the case
90.
by thimabi 1y ago
But what’s the future in terms of profitability of LLM providers? As long as features like Study Mode are little more than creative prompting, any provider will eventually be able to offer them and offer token-based charging.

More ›