Someone could turn this into an app. Scan the medical bill, analyze the codes, understand the rules, show a target to negotiate. Could be extra spicy by providing the public data on hospital pricing before arrival.
If they are going to lose the LLM race, at least they could invest in building a great frontend to interact with the non-MS models to capture a share of the revenue... Right now their frontend is one of the worst AND their model is one of the worst.
Nice strawman. The core of webapis is about opening up lower level functionality from the sandbox/accessibility of the web. Beyond audio and video IO, there's great stuff coming with webgpu and webNN. Web apps are much safer and much more convenient than downloading an app, well in theory they could be if support wasn't regularly sabotaged to protect a corporate interest in walled gardens.
Breaking things is not extending battery life. Battery life assumes functionality. Breaking functionality to extend it is a scapegoat and the break-whatever-you-want could be provided as a mode instead of one-size fits all, we don't care what breaks approach.
Or just the way the menus are on apps. Some app implement their own file/edit/view menus at the top of the app, then some will use the apple version at the top of the OS. If you plug in a TV to use as a monitor and cannot adjust the aspect ratio you're forced to blindly activate these menus as they're clipped from the screen.
MacOS folder navigation is a complete pain too, sometimes you see the list of OS folders, sometimes you see only the folder you opened in finder. If the menu is clipped due to the above aspect ratio problem, good luck getting to your home folder... No functionality to easily open a folder in terminal. Lots of basics just counter-intuitive.
Adds hours of battery life to the expense of making your microphone input completely inaudible due to throttling if you background the tab it's running on.
On iOS you cannot even keep a web app running in the background. The second they mutlitask, even with an audio/microphone active, Apple kills it. Are they truly adding battery life or are they cheating by creating restrictions that prevent apps from working?
Being able to conduct a voice call through the browser seems like a pretty basic use case to me.
If we dismiss remote management as a non-core feature shouldn't we consider installing a new browser to be advanced usage as well?
I understand that this post is about MacOS, but yes, we are forced to support Safari for iOS. Many of these corporate decisions to prevent web apps from functioning properly spill over from MacOS Safari to iOS Safari.
The basics are not rock solid. Even a core feature such as remote management crashes and freezes every 5 minutes when you connect from a non-apple machine, many have reported this over years but Apple just does Apple. Safari is still atrocious when it comes to web api supports. The worst part is, with Apple, we do not know if these are intentional anti-competitive barriers or actual software bugs. I purchased a mac mini simply to compile apps via xcode and can say the core experience is MUCH more buggy than a fresh Windows or Ubuntu install.
Edit: Hard to call intentionally preventing support for web apis a power user thing. This creates more friction for basic users trying to use any web app.
Edit2: lol Apple PR must be all over this, went from +5 to -1 in a single refresh. Flagged for even criticizing what they intentionally break.
Dietary fiber adaptation probably has many more benefits beyond liver health too. Gut microbiomes resulting from high sugar diets likely cause the majority of disease in modern times.
As I said sprinkle a bit of benchmarks polluting the training and you have your loop. Each iteration will be better at benchmarks if that's the goal and that goal/context reinforces.
> I'm sure some of it could be in the parallelization of processing that has to occur to service the large amount of requests. and more and more traffic are spreading it thin?
Even if this is the case, benchmarks should be done at scale too if the models suffer from symptoms of scale. Otherwise the benchmarks are just a lie unless you have access to an unconstrained version of the model.
The benchmarks are not typically ongoing, we do not often see comparisons between week 1 and week 8. Sprinkle a bit of training on the benchmarks in and you can ensure higher scores for the next model. A perfect scam loop to keep the people happy until they wise up.
More like churning benchmarks... Release new model at max power, get all the benchmark glory, silently reduce model capability in the following weeks, repeat by releasing newer, smarter model.
I really hope benchmarking improves soon to monitor the model in the weeks following the announcement. It really seems like these companies introduce a new "buffed" model and then slowly nerf the intelligence through optimizations.
If we saw task performance week 1 vs week 8 on benchmarks, this would at least give us more insight into the loop here. In an environment lacking true progress a company could surely "show" it with this strategy.
I thought they were just nerfing the models with optimizations shortly after public release benchmarks are released but it seems new "safety" infrastructure went live for both Anthropic and OpenAI that is causing all sorts of issues with routing to downstream models.
Anecdotally, I've observed both Sonnet4 and GPT5 behaving equally bad with code and sharing similar hallucinations from fresh chats. Is some sort of cross-company safety router akin to the great firewall being rolled out for AI chats?