Twitter still does have quite a lot of unique content that either appears there first or isnt accessible anywhere else at all, unlike paid article websites, previews without logging in actually work for the most part, and xcancel as you said is a thing. Which extension are you using for redirects?
For the best quality reply, I used the Gemma-4 31B UD-Q8_K_XL quant with Unsloth Studio to summarize the URL with web search. It produced 4.9 tok/s (including web search) on an MacBook Pro M1 Max with 64GB.
Here an excerpt of it's own words:
Unsloth Dynamic 2.0 Quantization
Dynamic 2.0 is not just a "bit-reduction" but an intelligent, per-layer optimization strategy.
- Selective Layer Quantization: Instead of making every layer 4-bit, Dynamic 2.0 analyzes every single layer and selectively adjusts the quantization type. Some critical layers may be kept at higher precision, while less critical layers are compressed more.
- Model-Specific Tailoring: The quantization scheme is custom-built for each model. For example, the layers selected for quantization in Gemma 3 are completely different from those in Llama 4.
- High-Quality Calibration: They use a hand-curated calibration dataset of >1.5M tokens specifically designed to enhance conversational chat performance, rather than just optimizing for Wikipedia-style text.
- Architecture Agnostic: While previous versions were mostly effective for MoE (Mixture of Experts) models, Dynamic 2.0 works for all architectures (both MoE and non-MoE).
I wrote it originally because I wanted my openclaw install to talk to my assistant's openclaw, and my openclaws that were local at different houses.
It's morphed a lot since then, and is close to being super useful -- it allows group chat, and is close to having a realistic API call on threshold vote gateway system built in.
That stuff is built to support Corpo's main business model which is providing real world asset and governance access to agents.
So, for example, I think agents might like to vote on sending a wire transfer by approving a specific mercury bank API call.
I could go on. You can also use it to remotely chat to an agent across firewalls - it's pull / poll only.
i think its all about caring and knowing what you want to make and willing to iterate on the result until it is actually good. If you want the ai to do your job for you its probably not going to work, but if youre really good at using its advantages you almost certainly will be winning
Yeah and also they still want to get at least some sales on the mac studios and mac pros with ultra chips, 256gb m5 max wouldve straight up killed both of those products.
We can have nice things but nobody is going to hurt themselves to give out things that are the very best possible, theres probably a lesson in this
> 256gb m5 max would've straight up killed both of those products.
1) Not necessarily, as the thermals would presumably be different, the use-case is different (not everyone wants or needs a laptop; expandability of the Pro, etc.) and Max =/= Ultra, especially if you're crunching local inference.
2) Even if there was some cannibalisation, does that matter? Unless we assume Apple is running a higher profit margin on Studio/Pro machines (unlikely, since laptops are more expensive than the equivalent Mini/Studio) they're still making roughly the same money at the end of the day. And for the higher end (i.e. workloads needing the Ultra and/or >256GB RAM) there's still no competition.
3) I'd not be surprised (RAM shortages aside) to see the RAM options on the Ultra increase before long, maintaining the differentiation, just at a higher level.
Basically, Apple stumbled into relevance as (amazingly) the most cost-effective option for local inference. Having found themselves in this position, it would be a huge fail to not lean further into this. They seem to be doing this to an extent by optimising chips for e.g. prompt processing, but increasing the RAM is needed too.
laptops seem kinda solved now? for every cpu vendor and every os theres now a great option that just, works, so everyone can just use what they like and not be at a disadvantage due to not knowing the latest developements in the space, or due to habitual preference. Good type of boring
Twitter still does have quite a lot of unique content that either appears there first or isnt accessible anywhere else at all, unlike paid article websites, previews without logging in actually work for the most part, and xcancel as you said is a thing. Which extension are you using for redirects?
reply