Agreed. I find that particularly annoying, and I also seem to find that the spatial arrangement or stereo effect is muted for most instruments (or the model simply doesn't use that feature as well as a good human musician).
I'm a psychiatry resident who has been into ML since... at least 2017. I even contemplated leaving medicine for it in 2022 and studied for that, before realizing that I'd never become employable (because I could already tell the models were getting faster than I am).
You would be sorely mistaken to think I'm utterly uninformed about LLM-research, even if I would never dare to claim to be a domain expert.
I'm still impressed by the progress in interpretability, I remember being quite pessimistic that we'd achieve even what we have today (and I recall that being the consensus in ML researchers at the time). In other words, while capabilities have advanced at about the pace I expected from the GPT-2/3 days, mechanistic interpretability has advanced even faster than I'd hoped for (in some ways, we are very far from completely understanding the ways LLMs work).
I'm a psychiatry resident who finds LLM research fascinating because of how strongly it reminds me of our efforts to understand the human brain/mind.
I dare say that in some ways, we understand LLMs better than humans, or at least the interpretability tools are now superior. Awkward place to be, but an interesting one.
I would be curious to know more precise numbers. My intuition suggests that when Sony sells millions of them, the number diverted for non-gaming purposes is maybe thousands or tens of thousands.
Given that Gemini 3 Pro already did solid on that test, what exactly did they improve? Why would they bother?
I double checked and tested on AI Studio, since you can still access the previous model there:
>You should drive.
>If you walk there, your car will stay behind, and you won't be able to wash it.
Thinking models consistently get it correct and did when the test was brand new (like a week or two ago). It is the opposite of surprising that a new thinking model continues getting it correct, unless the competitors had a time machine.
Previous models from competitors usually got that correct, and the reasoning versions almost always did.
This kind of reflexive criticism isn't helpful, it's closer to a fully generalized counter-argument against LLM progress, whereas it's obvious to anyone that models today can do things they couldn't do six months ago, let alone 2 years back.
Same here. I did notice what I think was an actual error on someone's part, there was a chart in the files comparing black to white IQ distributions, and well, just look at it:
>Interactive Human Simulator is a bold way to describe spinning up a few GPT calls with mood sliders, but sure, let’s call it anthropology. Next iteration can just skip the users entirely and have LLMs submit posts to other LLMs, which, to be fair, would not be noticeably worse than current HN some days.
>If anything the agentic wave is showing that the chat interfaces are better off hidden behind stricter user interface paradigms.
I'm not sure that claim is justified. The primary agentic use case today is code generation, and the target demographic is used to IDEs/code editors.
While that's probably a good chunk of total token usage, it's not representative of the average user's needs or desires. I strongly doubt that the chat interface would have become so ubiquitous if it didn't have merit.
Even for more general agentic use, a chat interface allows the user the convenience of typing or dictating messages. And it's trivially bundled with audio-to-audio or video-to-video, the former already being common.
I expect that even in the future, if/when richer modalities become standard (and the models can produce video in real-time), most people will be consuming their outputs as text. It's simply more convenient for most use-cases.
I found it genuinely impressive how useless their "GPTs" were.
Of course, part of it was due to the fact that the out-of-the-box models became so competent that there was no need for a customized model, especially when customization boiled down to barely more than some kind of custom system prompt and hidden instructions. I get the impression that's the same reason their fine-tuning services never took off either, since it was easier to just load necessary information into the context window of a standard instance.
Edit: In all fairness, this was before most tool use, connectors or MCP. I am at least open to the idea that these might allow for a reasonable value add, but I'm still skeptical.
1. I do know what to do with it. I take notes, a lot, in my work as a doctor. That's been the case since I owned an iPad Air from 2020, which I replaced with an 11 inch M1 iPad Pro (which broke), and I finally caved and bought a 13" iPad Pro to replace it. I ended up getting the M4 model because there just didn't seem to be older ones reasonably available. Even the M1 was more than fast enough for the overwhelming majority of iPadOS applicantions.
Why an iPad? Android tablets have been... not great for a long time. The pencil is very handy, and the ecosystem has the best apps. Also, I know a few rather handy tricks Safari can do, such as exporting entire webpages as PDF after a full-screen screenshot, that are very useful to my workflow.
2. I already own multiple general purpose computers. They're not as convenient as an iPad. My ridiculously powerful PC or even my decent laptop doesn't allow the same workflow. However, that's not an intentional software limitation, it's a consequence of their form factor, so I can't hold Microsoft to blame. On the other hand,Apple could easily make an iPad equivalent to a MacBook by getting out of the way.
3. The inability/difficulty of side-loading apps, the restriction to a locked down store. Refusing to make an interface that would allow for laptop-equivalent usage with an external/Bluetooth m+k. You can use an external monitor, but a 13" screen should already be perfectly good if window management and M+K usage wasn't subpar. Macs and iPads have near identical chips (the differences between an M chip for either are minor), and just being able to run MacOs apps on device would be very handy. Apple has allowed for developer opt-out emulation of iOS and iPadOS apps on Mac for a while now, why not the other way around?
If not obvious from the fact that I'm commenting on HN, I would gain utility from terminal access, the ability to compile and run apps on device, a better filesystem etc. Apple doesn't allow x86 emulators, nor can I just install Proton or Wine. If I can't side-load on a whim, it's not a general purpose computer. I can't use a browser that isn't just reskinned Safari, which rules out a great deal of obvious utility. There are a whole host of possible classes of apps, such as a torrent manager, which are allowed on other platforms but not on iPadOS. It's bullshit.
My pc and laptop simply aren't as convenient for the things I need an iPad for, and they can't be. On the other hand, my iPad could easily do many things I rely on a PC for, if Apple would get out of the way. iPadOS 26 is a step in the right direction, but there's dozens left to go.
I own an M4 iPad Pro and can't figure out what to do with even a fraction of the horsepower, given iPadOS's limitations. The rumors about an upcoming touchscreen Mac are interesting, perhaps Apple will deign to make their ridiculously overpowered SOCs usable for general purpose computing. A man can dream..
Daniel Kokotajlo released the (excellent) 2021 forecast. He was then hired by OpenAI, and not at liberty to speak freely, until he quit in 2024. He's part of the team making this forecast.
The others include:
Eli Lifland, a superforecaster who is ranked first on RAND’s Forecasting initiative. You can read more about him and his forecasting team here. He cofounded and advises AI Digest and co-created TextAttack, an adversarial attack framework for language models.
Jonas Vollmer, a VC at Macroscopic Ventures, which has done its own, more practical form of successful AI forecasting: they made an early stage investment in Anthropic, now worth $60 billion.
Thomas Larsen, the former executive director of the Center for AI Policy, a group which advises policymakers on both sides of the aisle.
Romeo Dean, a leader of Harvard’s AI Safety Student Team and budding expert in AI hardware.