vkaku

L take.

Honestly, the folks writing inference drivers for Llama.cpp / GGML would sort of benefit from better compiler work like this.

In general, Valve's work has been exceptional and supplementing AMD's own ROCM/OpenCL and Vulkan team, they've gotten a lot of defaults right and people should work together with Valve to improve support for their chips.

I really think the authors in FT are not considering:

- The innate value of the models come from not training chips nor hosting, but the actual quality of thinking traces, inference costs and usage. Anthropic shines here, but not for long.

- The variety of skills and evals on new tiny open models proves Claude is expensive and not extremely better than these models today, and math benchmarks alone do not cut it.

- Enterprises are realizing that they can just buy some regular Macs and do their own inference far cheaper than invest in Claude Subscriptions, and Claude given tools are losing their edge to harnesses such as Pi and Opencode.

- Nobody needs Hacking Tools unless they are red-teaming their own code. The niche of people wanting this could write their own tools and workflows to test the effectiveness of many LLMs on these workflows and see how far they can push it.

- The amount of collusions, hypes and fear mongering do not sit well with many people (including myself) - stick to the science and make usage safe, like everything else we do with software in production.

- The IPOs often get priced down with market discovery and anyone who's buying it ONLY is interested in the long run, not inflated valuations not hype not any of these lies used to sell this stock. I am bearish about the IPO at this rate.

- For me, reasons for SpaceX money investments are NOT due to AI *That part is a true waste of money* but good enough due to manufacturing, supply chains and actually building useful Satellite infrastructure; I still am bearish about them launching NVL72 nodes in space, not because of the space part - I just don't trust those chips enough for it to sustain that long.

Of course it's a sponsored post.

LLMs were a bad idea. MCPs were the band-aid to get them closer to the real world and do something for recursive self improvement. Then LLMs became better suited for everyday use.

I'm tired of these kind of posts. This author is on my black pill list.

It's great they do C23. I'd love to get three-four projects worth of things merged together for a good reason:

- TinyCC, SDCC, Musl and Cosmopolitan Libc

Reason is that I require a quick fast portable compiler with a decent runtime that can also be embedded into a larger runtime, like the way Bun does - while keeping runtime small and portable.

SDCC does not need the same amount of bloat - having decent small C runtime is a job in itself. Fixing it once could allow a better compiler that isn't quite as bloated as LLVM nor as small like TinyCC.

That gap isn't solved.

Those Razr phone updates need to start landing sooner ....

Many developers will likely be developing at Android 16 APIs only. Will rely on Aptoide+AOSP exclusively. This is likely the timeline we'll see a decline in Play Store and Pixel usage.

We all know these aren't purely technical reasons, it's to break the monopoly of one Internet Company's decoder format portfolio. You can't let one Fruit Company's format be the thing either, so it's basically a L for engineering but a W for pleasing everyone TBF.

Market correction will happen. Banks go down and close during recessions. Hope these people are wise enough to see through these effects.

Don't buy this argument. It's like saying, we lords who hold this capital will build all this economically destructive stuff anyway, and wait for the world to not react to our stupid ways of enriching yourselves.

People are not going to slow down because this was brought to them on less than endearing terms. They don't see any of these stated noble intentions.

Claude was already used to cause economic, political and social destruction. As Anthropic is seen doing it, others aren't going to just sit down, read the blog and say, oh, I'll stop developing models because Dario, you touched my heart with your true words.

Here. You are not alone. Just run it and host models yourself.

https://github.com/guilt/xet-server

Here: https://github.com/guilt/xet-server

This is great. All algorithmic efficiencies are amazing!

One thing I'd remind all scientists and the wonderful people here is this wonderful meme/line from Jurassic Park: "Your scientists were so preoccupied with whether they could they didn't stop to think if they should."

What is the actual amount of data that needs to be pre-trained and what is not? Nobody has come up with great answers to this question, and I'm already seeing amazing 0.5b-2b parameter models working very well with n-Gram corpuses of data. So, how many parameters do you really need for a given workload?

Next Quarter? :) That's too long

LOL ... Accurate ;)

A lot of time accessibility is about that extra effort one can't get with tight development budgets. I apply generous amounts of butter while I keep the thing dry and let it be otherwise accessible at the same time.

A lot of the AI progress is stalled because these chips do not handle thermal contraction/expansion well at all, and end up permanently destroying these chips. HBM has been a failure at this scale with chips getting destroyed every 4 months or so during operations.

I'd like to see some actual science saying, here was the problem, here's how we solved it, here's the AFR data, here's this running after X cycles etc. Nobody has done this reliably yet. That entire industry is hiding the bodies.

I'd like to next the article which says almost no need for training models. That's really what I'm interested in at this point

It's a good step in the right direction. Hardware reliability and testing needs to happen. Tell me how long it can run without heating up.

Oh, look, more state sponsored control propaganda disguised as protection.

Scams are run by state sponsored actors and intelligence agencies and their targets are often regular people who have nothing to do with any of this.

You can say whatever you want, DNS is the only thing preventing control by all governments. If that locks down, you find see more people pushing for alt root servers and other ways to get out of there. What might actually help is tracking finances of government agencies and the few people who are often funding all these illicit activities.

Ill-devised Internet control movements will make everything incredibly difficult to reconcile back to. Parts of the world are already creating and using their own systems for this, and further DNS control will lead to even wider adoption of Alternative DNS Roots such as OpenNIC or Handshake even. (See https://en.wikipedia.org/wiki/Alternative_DNS_root )

Eventually, given that and a list of alternative certificate authorities, we will end up with a permanently forked internet where the only saving grace might be the protocols themselves and nothing more. Some companies are already creating a way to send a list of trusted CAs via their DHCP infrastructure already (See https://info.support.huawei.com/info-finder/encyclopedia/en/... ) so I can totally see loss of control happening in the upcoming years.

You heard it here first.

Issue with storage is you need shelf stability without ongoing electrical/mechanical work refreshing data time to time. Masked ROMs and M-Discs are those two things which can really last a very long time; Floppies, Hard Disks, Solid State Storage and present day ROMs all require periodic costs, efforts and refreshes to keep them going (that means they can't be kept in storage and then re-opened in 30 years).

We also need a mechanism that outlasts that archival lifespan so how to read Discs or Masked ROMs and how to interpret them, also needs to be stored as well. Repairing electronics is a valuable skill that needs to be passed on across generations as well.

Keeping humans around seems to be the best bet, given how things are going, I wouldn't count on the present set of beings around to preserve any of this information that well either.

I am going to either buy a CommodoreOS phone (Jolla) or a Linux phone (TBD) at this point for any such purchases;

I currently have old phones who do not have this restriction, but if Google Play did this, I'd already be ready with Aptoid and ready to disable Play Store if I need to.

I already got a working Xet server with fully efficient `hf upload/hf download` functionality. I think I don't need much honestly than to mirror/proxy all the models and datasets I need and get some simple repo clones to exist.

If there's interest I'll post an article on how to not let ONE company keep access to models.

I'm giving out more unpopular opinions today for entertainment and thought :)

Every model is benchmaxxing today. Many of them are awful until tuned for any particular task.

I don't think personal computers will be a BIG thing ....

- IDK, some random mainframe company exec in the 60s ;)

Many people seemingly also use these tools against their data center vendors so I believe this is just for consumer users?

Supporting old games is important for me or I'll use a distro that doesn't fuck around for me.

Microsoft fucked up many people's lives and I for one will never let working products get fucked with the same terrible people and their decisions.

This is great. It tells you how many manufacturers make it easy and difficult to get refunds. ASUS is the most specific one, albeit their refund range is between 9 and 65 EUR. Great work.

We really need an artifactory for GGUFs

If this were I I'd rip every CD and upload FLACs to the Internet Archive, vibe code a free music playing app and release it for free.

None of these restrictive arguments would stop me from doing the next 2-3 steps even, after buying those many discs.