RetroFed
eicker, eicker@lemmy.world
Instance: lemmy.world
Joined: 3 weeks ago
Posts: 18
Comments: 166
Hi! š Iām Gerrit and in love with the Internet since 1994. š Media ā¹ Communication šØāš¬ Entrepreneur & Consultant at eickerĀ® [iCare] ā» Reinschauen: eicker.TV š² Get in touch! šš©šŖš¬š§ #Fedizen #AI #Tech #Media #Sovereignty
Posts and Comments by eicker, eicker@lemmy.world
technology
europe
Comments by eicker, eicker@lemmy.world
The industry keeps benchmarking against other labs instead of against user needs: If a 3.5GB model answers 95% of everyday questions well enough, the remaining few percent has to justify hundreds of gigabytes of weights, huge energy bills and constant cloud costs.
Stock prices arenāt proof of being right, but they do show investors can change their minds a lot faster than the narratives do.
Appleās marketing deserves skepticism, but the technical argument is separate. Local inference does not eliminate giant training clusters, it mainly cuts inference costs, latency and improves privacy. Apple still uses cloud models when needed.
Cookās biggest product might be expectation management. He rarely promises tomorrowās miracle, which buys Apple room to ship when it suits them instead of when Wall Street gets impatient.
Apple has always been unusually willing to sacrifice short term hype for long term positioning. That does not guarantee they are right, but it is a very different bet from spending hundreds of billions assuming demand will eventually justify the buildout. If AI demand disappoints, discipline suddenly looks a lot more valuable than scale.
About 4+ maxed M4 Studios, I guess. But thatās not the point: in 80%+ of cases, people wonāt need that kind of AI model to solve their problems.
Raw capability is only one metric: A local model probably will not beat the best cloud model any time soon, but it does not need to. If it handles 80 to 90% of everyday tasks instantly, privately and at near zero marginal cost, that is a huge win. Reserve the cloud for the genuinely hard requests, not every prompt.
Heheā¦ š
Absolutely. Looking forward to seeing the next generation of Macs.
The decentralised operation of LLMs would also be significantly simpler and cheaper for the use of decentralised renewable energy sources.
Iām waiting for the next generation of Mac Mini and Mac Studio.
The present: Open Weight AI, such as Kimiās, is already almost exactly as good as ClosedAI from Anthropic and Ā»OpenAIĀ«.
It would seem so. On the other hand, it is puzzling that they did not also allocate the necessary resources to the development of LLMs. š¤·
The interesting part is not whether Apple wins the biggest model race, but whether it changes the economics: If enough AI runs locally, every token avoided is cloud capacity nobody has to build. That is a very different business model from selling ever more cloud compute.
Itās easy to underestimate how different these facilities are: Traditional data centres are built to run millions of separate workloads efficiently. AI data centres are engineered to make tens of thousands of GPUs act like one giant supercomputer. The building may look similar, but the power, cooling, networking and economics are almost an entirely different industry.
The market is fragmented across the Member States, each of which conducts its own research. That is precisely what needs to change. Hopefully, this approach will work.
This is a significant step for Europeās AI ambitions. Access to large scale compute has become one of the biggest barriers to building competitive AI models. If executed well, these gigafactories could give European startups, researchers, and companies the infrastructure they need to innovate at home instead of relying almost entirely on US or Chinese providers.
This is exactly how the DSA is supposed to work: Once a platform reaches sufficient scale and societal impact, it faces higher transparency and accountability requirements regardless of whether itās social media, gaming, or AI. If anything, it would be more surprising if major AI platforms werenāt eventually included.
These are purely paper profits.
Unfortunately, it is highly likely to take decades before Meta loses its relevance: the network effects are enormous, and the vast majority of people over 35 will probably only leave Instagram and (from the age of 60) Facebook completely in exceptional circumstances. And I donāt even want to get started on WhatsAppā¦
Itās not about who deserves to win, itās about incentives. If one side unilaterally slows development while others donāt, the likely outcome is a shift in technological leadership, not a slower pace overall. Whether thatās the US, China, or anyone else is a separate political question from whether voluntary restraint is an effective strategy.
Ā»Slow down AIĀ« sounds noble until you ask who actually slows down: Democracies and companies that comply simply hand the advantage to rivals that donāt. History suggests voluntary technological restraint rarely wins against geopolitical competition. The race doesnāt stop: it just changes leaders.
Pacing the Frontier: A statement from 1,132 employees of frontier AI companies (pacingthefrontier.com)
This is why media ownership matters: When the people deciding what gets amplified are also celebrating political power brokers, it becomes harder to separate journalism from influence. Even if Politicoās newsroom stays independent, the optics alone are enough to make readers question where the lines really are.
Exactly. Platform loyalty is a bad archiving strategy. If one platform disappears, changes the rules, or buries old content, your work and links still survive instead of vanishing with someone elseās business decisions.
The irony is that this argument will probably get more attention on Reddit than on any open forum: Centralised platforms are great for discovery and awful for preservation. Use them as the billboard, not the library. Keep the valuable stuff where you control the rules, the backups and the future.
A $250 billion backstop is an extraordinary amount of financial engineering for an industry thatās supposed to be the next industrial revolution: If AI demand is as inevitable as claimed, why does every new data centre seem to require increasingly creative financing and ever larger guarantees?
Nvidia and OpenAI in talks for up to $250 billion dollar backstop to fund AI infrastructure plans (cnbc.com)
Because it was always so wonderfully easy to do nothing. Costs have risen, and new realities are increasingly being created in both politics and the economy: in France, Denmark, Schleswig-Holstein and so many other places, large and small.
Without a doubt. And yet: without open source, pretty much everything digital would come to an end straight away. Sometimes whatās visible isnāt what really matters.
Weāve spent years telling kids that every problem is solved by putting another screen in front of them. Now weāre discovering that unlimited access to the worldās biggest distraction might not have been the revolutionary education strategy we thought it was. Technology is a tool, not a teaching method.
Ever since facts have been replaced by ideologies, the obvious has increasingly become a mystery.
Because digital sovereignty isnāt a job title, itās a procurement strategy: The hiring shows up as cloud engineers, cybersecurity, Linux, semiconductor, networking, public sector IT and compliance roles. The money is real, but it mostly flows through existing industries rather than creating a shiny new profession.
Absolutely. Trust isnāt binary, and it isnāt unique to Chinese models either. Every (frontier) model reflects the incentives, values and constraints of whoever built it. Thatās exactly why competition matters.
Unfortunately, based on many years of experience, I have to agree. However, I also see a light at the end of the tunnel, particularly in Europe, where there is a growing desire to break free from the US SaaS stranglehold, using open source software.
Regarding OWAI I see a fundamental difference: Slack isnāt just software, itās a hosted service with identity, storage and network effects.
An open weight model is more like a compiler: once downloaded, nobody can revoke it. You may still pay for inference, but pricing power drops when anyone can host the same model.
Funny how Ā»open is dangerousĀ« only became the dominant narrative once Chinese models caught up: If your competitive moat depends on governments blocking free alternatives instead of building better products, maybe your moat isnāt technology at all.
Exactly. Itās much easier to regulate an API than software people can download and run on their own hardware.
OnceSince capable open weight AI models exist, they spread much like Linux or other open source software: mirrors appear, forks emerge, and improvements compound. Governments can make access less convenient, but preventing global distribution altogether is a far harder problem than restricting a hosted service.Thoughts on "Open" Models: The debate focuses on balancing open and closed models, with cost and geopolitical considerations influencing the push for open weight AI. (spyglass.org)
AI companies keep promising productivity gains, but one of the first guaranteed outcomes is that everyone else helps pay the electricity bill. If the grid is upgraded for a few hyperscalers while the costs are spread across millions of customers, thatās not a free market. Itās a subsidy with extra steps.
The weird part is that everyone keeps calling it public infrastructure, but if the primary beneficiary is a handful of private AI companies, that argument starts looking pretty thin. Reliable grids benefit everyone. Dedicated power corridors for hyperscalers are a much harder sell, especially when locals bear the costs.