Nokia 4.2

This post just gives me more questions than answers and I'm unable to form a decision: Why was v3.4.1 the most buggy, right before the Claude commits? [...] Why does v3.4.2 have 0 bugs, or 0 bug score. And why was such an outlier (no other commit seemingly has this??) allowed to mix into aggregate statistics and bring all the "is Claude buggy?" scores down. [...]

Why was v3.4.1 the most buggy, right before the Claude commits? [...] Why does v3.4.2 have 0 bugs, or 0 bug score. And why was such an outlier (no other commit seemingly has this??) allowed to mix into aggregate statistics and bring all the "is Claude buggy?" scores down. My original metrics which didn't filter out feature requests and questions had it at four bugs and prior to that it was even higher and it didn't make much of a difference to the overall analysis (fell well within the IQR, the lower end of it too). [...]

[...] We went from some commits every so often (because rsync was basically finished software) with a handful of issues a month, to "a bajillion Claude commits then a release" with a pile of issues within weeks, at least one of which is about how the goddamn thing doesn't build. [...]

If I invest half my income and spend half my income, and the prices of goods goes up 4.2% and my income goes up 4.2%, then I've made progress; I'm now investing more than half my income, because the half of my income I was spending has stayed even and the half I was investing has increased. [...] The prices of goods goes up 4.2%, so to keep your current living standards, you must now spend $51.20. [...]

[...] Gemini on the other hand messes up more frequently and sometimes does weird things like try to delete files it’s not familiar with, at least the 3.6 flash model I’ve been using lately does this. [...]

Funny enough, I have a simple test that I have been running on each new model that catches my eye on OpenRouter. It is just a short prompt that asks the model to research yesterdays news and summarize it in a specific format along with a critique of the article or a highlight of any bias: I have been doing this test for about a year now. I run it on different harnesses and apps as well just to give me an idea of what they differences between them might be. Since I have been running it for a while now, I have a good sense of the correlation between this output and how the model will be on the rest of the things I want it to do. I would say that almost every model (just tested Mercury 2.5 Preview and IBM Granite 4.2-8b) are pretty much the same on this task. Some are more diligent with how many sources they go out and get, but for the most part they are all very close in quality and will follow the instructions very well. So saying "everything struggles" with that has just simply not been my experience.

I recently upgraded an old website[1] stuck on Rails 4.2 / Ruby 2.4 (yeah, I know) to Rails 8.1 / Ruby 4.0 and it's still awesome as it was in the first days of Rails 2.3 back in the days. Such a great framework and language to work with! railsdiff.org was really helpful to upgrade in addition to the official documentation, especially now with LLMs that can process all these text and assist you in the process. [1]:

Donald Trump said “I love the inflation” after new data showed that inflation jumped to an annual rate of 4.2% in May, the third consecutive monthly increase since the start of the Iran war and a three-year high. You may be on to something...

Your comment is like 6 months to a year late. There for awhile it seemed like we’d have 3 big competitors but then Grok 4.2 or 4.4 was just diabolical while OAI and Claude continued their significant improvements. Grok was/is so bad that I was convinced musk was gonna shut it down and just fund Anthropic compute once they reached their compute agreement.

I used Google API / Yelp API to get top food reviewed places (4.2 star and 500+ reviews) and let used Claude API to review individual listing and menus twice. Google API has free API offered when you first start ($300) so you can try it out!

I wonder how you got it approved; it seems like a direct violation of guideline 4.2: 4.2 Minimum Functionality Your app should include features, content, and UI that elevate it beyond a repackaged website. 4.2.2 Other than catalogs, apps shouldn’t primarily be marketing materials, advertisements, web clippings, content aggregators, or a collection of links. Not that I would expect consistency from Apple reviewers, just noticing. And, while we are on this topic: maybe it indeed makes sense to make a website instead of an app? You can send push notifications the same way, but users won't need to install one more app and could use their favorite browser with their favorite settings, bookmarks, etc. etc.

I'm always wondering whether or not SDL would be better SDL supports mobile; GLFW doesn't. Maybe they're unifying codebases between the platforms. Edit: BTW SDL3 still works even on Android 4.2. I made an app using it and ImGui, without any Java.

FYI: When looking around my favorite spots for decent, cheap food in the small towns in Ohio where I hang out, it's easy to find plenty of them with a score of at least 4.2 on Google. [...] (And by that point: It's not cheap anymore.)

Consumer price index is about consumer goods. This is why tarrifs and such are considered regressive - they hit people harder the less money they have because a larger percentage of their spending is consumer goods. If I invest half my income and spend half my income, and the prices of goods goes up 4.2% and my income goes up 4.2%, then I've made progress; I'm now investing more than half my income, because the half of my income I was spending has stayed even and the half I was investing has increased.

The few things that make me agree with GP: 1. "AI" is a marketing term used by the likes of OpenAI/Anthropic/Google. LocalLLaMa communities prefer to use "LLM" or "model". So for a lot of people "AI" is just a service (see 4.) 2. "AI capability" is an irrelevant spec and marketing slug. The hardware specs will give you the needed infomation to consider a model[0][1]. 3. If you'll want to run a model locally, you'd know that a midrange notebook isn't the device to look for. Instead, look at workstations with discrete graphic cards + lots of VRAM (24GB+), Strix Halo APUs or a MacBook with lots of RAM, or some dedicated workstations like the NVIDIA DGX Spark[2]. 4. An inference engine can run anywhere, you can pick any LLM hosting service. LLM clients just expect an API endpoint anyway. [0]: [1]: [2]: