ezst

Reading about so many people daily-driving laptops built around a decade ago, if not more (which includes myself, happy owner of a T480s), and not suffering any major penalty in terms of overall usability and battery life, it really shows that the returns of having more CPU power/smaller lithography have been marginal for a very, very long time (CPUs haven't been an enabler for new use-cases/a major incentive to upgrade a laptop in recent history, AFAICT).

That's not to say there hasn't been progress in this space, but I'm also amused by the many comments along the lines of "my 2020 M1 is so revolutionary that it's all I need in 2026" in every other Apple thread. Probably all we've been needing has been some 14/10nm 8th/10th Gen Core i7 in a decent package? And about package, how come that this era's ThinkPads and pro Dells/HPs get a bigger "love-factor" than their contemporary intel MacBooks?

Now, reading more into it¹, it appears that what Intel was shipping 8th-generation Intel Core i5 8257Us (i.e. a 2017 Coffee Lake chip) in its MacBook Pros for as late as 2020, before switching over to the M series. No wonder they weren't getting much love!

¹: https://en.wikipedia.org/wiki/MacBook_Pro_(Intel-based)#Tech...

Or more probably, it's imputable to biology and the natural replacement of the "old guard" by adults who didn't grow up during nukes schizophrenia. Or in other words, I doubt that those folks reasoned themselves out of the irrational fear for nuclear energy that they did not reason themselves into.

Same stuff, is a Markov chain too much (Gen) AI already?

I haven't read TFA yet, but having studied academically and then spent years implementing ML techniques for industrial applications (in the areas of machine vision and signals processing), I think that the premise of "AI exemption" is just nonsensical, by nature of being ill-defined. AI isn't a singular product or technology, it's an abstract family of computational techniques, some more complex or mathematically involved than others. Finding function parameters to fit data points using a least squares method is ML, therefore AI, and you would find that in every thermostat, ABS system or electric bike. Too simple? Here is a gradient of increasingly sophisticated ML techniques, with LLMs not even being on the top of the list. Where do you place the cursors of "what's too much AI"?

My pixel lets me pick between 60 or 120Hz refresh rate, the bump in perceived smoothness is substantial, to the point that 60Hz does appear choppy in comparison.

What about OP's comment rings communistic, or even political to you? Or is it the link between information control and wealth consolidation that you find controversial?

There's no paradox, there never were agents to begin with. Those are programs. It's same old software. If you sell me a software to do X and you fail to deliver, it's on you.

Even less likely to work in 2026. The only reason why the west-aligned world does the bidding of Walt Disney & friends, and doesn't instead shrug at DMCA and other extraterritorial IP laws, is to stay in good trading terms with the US. Now consider how much of that incentive remains when you are on the receiving end of illegal and arbitrary tariffs.

Isn't that a project where they must implement under white room conditions (or risk copyright infringement claims), and are careful about not taking contributions from people having been exposed to windows' source code? IIRC that's the case of wine

Can't disagree more, but heh :-)

Thanks for the link. I did go through it despite the annoying style.

It postulates that semantic models delivered in areas where SQL couldn't, that what hinders their broader adoption is the user-interface/querying layer (i.e. too few people adopting SPARQL), and suggests that a new dialect of SQL could solve that, especially because LLMs are so good at producing SQL.

> Knowledge graphs and semantic models promised something enterprises desperately needed: governed meaning, reusable business logic, and explicit relationships across complex data landscapes. And technically, they delivered.

For me that's the main argument: what is the value proposition for knowledge graphs and semantic models when you already have an established relational datamodel? Some argue that the graph's fluid schema brings flexibility benefits, but looking at the problem in its entirety (considering the effort to define and formalise the model, the effort to store content, the effort to retrieve it, deliberately keeping the dev/sysops and tooling out of the way), what you spare at formalisation/insertion stages, you pay back at retrieval (or, if you don't validate or constrain your inputs, expect surprises as an output).

That's where I stand right now:

- convince me that relational models can't express as powerful constructs and concepts as semantic ones

- convince me that "fluidity" in the model don't come with more troubles down the road than benefits

> Knowledge graphs exist. […] And almost nobody queries them directly. Not because the models are wrong. Not because the semantics aren’t useful. But because accessing them requires SPARQL

Or, guess what, a similar outcome can be explained differently: perhaps it's not SPARQL that's too difficult to grasp. Perhaps the alleged benefits of semantic models just do not materialise in practice, and there is no point to engage with SPARQL to begin with because there's no payoff at the end of it. I think the data processing ecosystem is sufficiently vivid and diverse to demonstrate that users have no problem embracing new technologies and languages as the need arise (sed/grep/…, hadoop/spark, pandas/polars/duckdb, decades of ETLs in graphical/textual forms, …)

Then the article is full of weird takes, especially within the "Hidden Cost" section, that I don't feel like commenting. Nor why their "SQL with semantic ontologies" is just syntactic sugar to save your editor from writing joins on your behalf?

Almost like, mercurial had it figured out and got completely ignored because of the dominance of git's monoculture... The unit of work to be captured is the series (of commits), there's no issue having a linear history as long as series' boundaries can be identified. Merges can fit that bill, but doing it right and consistently with git is real work and consequently nobody does it.

Out of curiosity, have you looked into capture checking from Scala? It's proposed as a generalization of the rust borrow checker, to handle capabilities and other side-effectful computations (error handling, asynchrony, ...) in an ergonomic manner. Not sure if/how much cross-pollination there might be :-)

https://docs.scala-lang.org/scala3/reference/experimental/ca...

Strikingly similar story here. Ontologies and knowledge graphs seemingly make a comeback because it is argued (with little evidence, from what I can tell) that it helps LLMs make sense of the dataset. I see no practical difference or benefit vs. providing the date model as DDL or via other means to represent the ER graph. I'd like very much to be convinced otherwise or at least read a rationale about the alleged differences.

you and me are enthusiasts, we love to know about codecs, formats, and we spontaneously seek novelty in this field and engage with it (I have similar stories of amazement about opus, I even went quite deep the rabbit hole of writing a java decoder implementation for it just to study it). But most people won't get nearly that excited, unless a new format magically enables something like twice the compression performance (so people start spreading the word that they can e.g. buy their next smartphone with 128GB storage and not the 256GB version and enjoy the same capabilities). Anything less than that will be a rounding error and barely perceptible, not worthy of the attention, and definitely not of the time to re-process and massively convert prior content.

Only if you believe this to be representative of a road legal car. I find it easier to believe that it's a purpose-built stunt prototype, otherwise they couldn't stop talking about how clever they were solving the seemingly infinite list of engineering challenges and contradictions that make it an actual car that actual people can drive on an actual road.

Having lived through the MP3 era and experienced the many years/decades? that it took for experts across fields to come together and figure out LAME and the bag of parameters that ultimately made it better than most audio formats around it, I would lean towards waiting this one out, and not take any announcement of a "winner" as gospel. We've yet to see the codecs mature and refine for some time.

I also remember the big claims about WebP's superiority, which ended-up in practice, at equivalent file-size, mostly trading away some of the JPEG artifacts that we all know for some unusual new ones.

I don't think there's a large-enough margin in this space that a new format could chew into to make the general public care: image compression seems like a solved problem overall. The second best thing is to come-up with a format that can make the content producers not have to care either, i.e. be the best in class or close-enough in a variety of use-cases (lossy, lossless, fast, animated, hdr, progressive, …). My naive impression was that jxl is mostly interesting in that regard.

Why not both? What makes EVs meaningful is when you get to the point of pushing green electricity to the grid and have the bulk of the utilitarian traffic consume it. Electrifying supercars for the sake of it has no measurable impact in the grand scheme of things.

> My impression also is that the Matrix community really loves their DAGs and the group state management that they built

I mean, if you take that out, what's left of Matrix and its differentiating proposition?

Not OP and loving that the upper part of that linked page comes with a warning message:

> Your account, @gmail.com, doesn’t have a recovery phone number. If you’re locked out of your account, a recovery phone number helps you get back in. [Add recovery phone]

I want to believe this, but I know Google's incompetence too well to doubt something like this is intentional :-)

As someone who came to terms with dying on the mercurial hill, someone who forgot more gitisms than there are to learn in the first place because git cli's inconsistencies and poor taste hurts and neuters my memory, yeah, I relate.

Most people aren't VCS nerds, most are of a generation that only ever knew git. Learning git was a prerequisite they engaged with out of necessity, not by choice, not after exploring the problem space and pondering about different approaches. Learning git was difficult, and the more they learned the more they discovered they had more to learn, including the little things, obscure details and a plethora of paper-cuts. For them (again, I suspect they are a majority), the perspective of learning something else comes with the expectation of a similar daunting effort and steep cognitive reluctance.

That's the paradox of git as I see it: had it been a simple and easy system to pick-up, it would probably have been replaced, and/or the VCS space would be much more competitive.

So it's mercurial all the way down, then?

Isn't that in parts why alcohol and tobacco are taxed at higher rates? That's a thing in many countries at least

What happened there, ... After official support by Google for the 4a stopped, I remember checking LOS out, and found that it wasn't supported there either. Did I hallucinate that, or did LOS resume support afterwards? Or maybe I'm mixing up LOS and graphene

And neither shows the outcome as to illustrate/support their point, which is phenomenally ironic considering the context..

> but there are holdouts still on Pandas because it was so complicated to learn with so many little extra rules to learn to avoid paper cuts that they feel like learning another data manipulation tool would be really hard.

Totally unrelated but I love and dread how that describes almost to the perfection the sad state of affairs we're at with git.

I see several problems with this,

- by definition, the AI isn't human, irrespective of its computational abilities, it needs a human to tell it what being human is, and the ways "humanity can be advanced" need to be validated on this basis

- "advancing humanity" isn't a one dimensional problem/single-KPI optimization game, you might optimize certain things (e.g. expected life expectancy) at the detriment of others (e.g. freedom of movement). If you are not understanding the proposition, you are not understanding the target outcome.

- even if the target outcome is perfectly formally specified and commonly understood (which it can't), you want your AI to rationalize that the journey to get there is the most direct and efficient.

- especially so since subsequent runs of the same prompt will provide different plans

- anything less than that is begging to be conned: as a sensible person, you wouldn't give unlimited power and all your faith to a single individual trusted to "advance humanity" if they cannot be understood. On what ground would you give the machine a pass?

In all, such comments really worry me. It's like AI is triggering for some people the kinds of oppressive religious feelings whereby the individual should submit itself to the will and desires of a pretended all-powerful being. Do you really think our ancestors chose to cut off their legs in abandonment when they figured that some animals could outpace them? No, they built traps and throwing weapons to catch them and see how they taste.

Not sure I get that. If AI is not working for humans' benefits (i.e. it is not working towards an optimization problem set-up by humans), whom is it supposed to benefit, then? (i.e. how is that better than an entropy-producing machine)?

Same for 2, there are so many infinite ways to boil the oceans, but so few oceans to boil to begin with. Better make sure that this insane energy (both in the physical, due to natural resources scarcity, as well as intellectual) is spent towards meaningful and useful ends. We can no longer be the judges of that if we can't comprehend what we got in return.