kettleballroll

What alternative can you recommend?

outside from very specialist professional software (AutoCAD and Photoshop come toind), I think this is mostly about getting over the hump of inertia. Both myself (software Dec and ai) and even my parents (browser machine) use Linux for ages without hickups.

I disagree. This article is typical embellishment from the university that produced the research. The reality is that they're likely pre-trial stage. Even if they make it through all stages of clinical trials (historically, chances of a drug candidate doing that are 2-5%), that usually takes the better part of a decade. So the answer is "no, we're very far from curing it".

Yes. The tldr is that this is a drug that showed promise that is in the early pretrial stages. Most drugs fail before they reach the market. Even if this one doesn't, it will take a decade at least before this will be available.

I needed to answer that question a month ago. The One felt too untested. Most people recommend the Flint2, but it felt expensive. So I landed on the asus RT AX53U

I thought the temperature only affects randomness at the end of the network (when turning embeddings back I to words using the softmax). It cannot influence routing, which is inherently influenced by which examples get batched together (ie, it might depend on other users of the system)

Typically in these tests you have three options "A is better", "B is better" or "they're equal/can't decide". So if 56% prefer O3 Mini, it's likely that way less than half prefer O1.also, the way I understand it, they're comparing a mini model with a large one.

Who if not Google was the first in generative ai? They invented transformers and diffusion, the cornerstones of text and image generati, respectively.

I'm a power user that games... so Windows it is. For now. The moment I don't need Windows to game, and Wine can run all my legacy apps, then I'm jumping ship to Linux.

As a power user that games on Linux, and has been for a decade -- what's stopping you right now? Which apps specifically tie you to Windows?

How did you find these things? Since you "stumbled upon them", you probably didn't know what you eee looking for, so... How did this research get started?

I don't just mean version control systems, but since you mentioned them: CVS (concurrent version system), rcs (revision control system) and subversion all seem fairly descriptive to me?

Same. I miss the old times when people tried naming their projects sensibly. I mean, we're constantly telling ourselves how variable and function names should speak for themselves, but then we name our projects using random, completely non-descript names. It's a annoying.

Are the book choices bad? I looked them up after first hearing about them through this article, and at least the Interviewing Users one sounded useful, and Principles seems to have a ton of good reviews. For the uninformed like me, could you give some color to those books?

I'm sorry, but I fail to see how your sources are relevant. [1] is written by a science fiction author, and while it outlines how palantir is trying to get into a contract with the NHS (apparently mostly by acquiring companies that already have such contracts), it doesn't say _anything_ about what palantir did (or has planned to do) that would be detrimental to the NHS. [2] merely speculated how palantir might help Ukraine defend itself from Russia. Which... Uhh... I don't see as a bad thing?

Are there any publications out there analyzing this more in depth? How are these datasets scheduled? Do you have your highest quality data first, or do you actually train using "dumb" data first until you establish some general language understanding before giving the high quality information? There is a lot of interesting research to do here that I'm sure people have already investigated....

Good for you if you can create a consulting business out of stating the obvious I suppose.

In my experience, tech problems are a lot easier to solve than people problems, and a lot of things that don't go well in a project turn out to be people problems. E.g. here are a few issues I encountered in my current project at work in the last month: "their framework makes assumptions that don't apply to our code, so we reimplemented the metrics instead of trying to integrate their version" or "the data was labelled wrongly, so we had to work around that", or "this coding convention is slowing us down". Once I tried digging down, it turns out they were all people problems in disguise, and they could all be solved by "stating the obvious". Do you never encounter issues on team / across teams, where in the end it turns out a lot of issues are just people not talking to each other or misunderstanding each other? If things are too hairy, I can definitely see the value in an external consultant helping disentangle these sort of problems.

I'd you're looking for an AI mentor, shoot me a mail (see my profile). I have over a decade of experience and work at one of the big AI labs.

I'm not sure HN isnthe right place to discuss political issues, but this is a horribly toothless law.

The law applies when a driver within a city is going 60-plus kilometers per hour over the speed limit. Outside the city, the vehicle must be going 70 or more kph (44 mph) over the limit. But at that level, the vehicle is only confiscated for two weeks. For repeat offenders, as well as for those who go 80-plus kph (50 mph) over the speed limit in a city or 90-plus kph (56 mph) outside a city, the vehicle is permanently confiscated and sold.

Doing 110 kpm in a 50s zone is willingly putting people's life at risk. It's beyond reckless driving, and might be considered as a super weak form of attempted murder (and suicide) even. And you get your car taken away for 2 weeks. Unless you're doing the Autobahn max Speed in a 50s zone or be a repeat offender for it to actually trigger.. What the hell?

Even better, if the racer doesn't own the car, there's (as they admit themselves) nothing they can do.

Why is this so weak? Is there some ethical concern about taking someone's car, or is this a political maneuver?

As others have said, it depends. What is weird to me is that the citation doesn't state whether these side effects are permanent. I also had LASIK, and did experience severe dry eye afterwards, but it went away within a month or two. I still have halos/bad vision at night, but I don't consider it life-altering. I would have the operation again in a heart beat.

As an aside

His work has been featured everywhere from Scientific American to the Lex Fridman podcast and The New Yorker.

This is a weird way to posit someone's scientific achievements. Had they said eg Lancet, Nature and Science -- ok, clearly someone publishing in those venues is a scientific heavyweight. But being featured in pop-science, a famous podcast and a general audience magazine only tells me how well someone can explain/sell their research, but doesn't actually say anything about the strength of that research.

The entire point of this is that productivity does not go down in a 4 day workweek

Just an anecdote, but: my productivity definitely took a huge hit when I reduced my hours to 4days/week. All of the "working harder/more efficiently/more focused" doesn't help if you already worked hard/efficiently/focused. If you have a job with a lot of slack/downtime/waiting time... Sure. But sometimes, working 20% less means that 20% less work gets done.

But the amount of work you're able to do is. And companies need to make money, they're not s charity. I took a 20% paycut to work only 4 days a week. And I think that is fair, because my output definitely did decrease by 20%. This might not be true for everyone, but all of the "working more efficiently" thing doesn't really help you if you were already working pretty efficiently 5 days a week.

I come from a very small provincial university, but they still reached out to me 2018, so I don't think so. (I bombed their interview and didn't get an offer, though).

But they're far behind in adoption in the AI space, while TPUs have both adoption (inside Google and on top) and a very strong software offering (Jax and TF)

N. Usually THE AUTHOR pays the journal to publish anything. However, the professors need publications for their CV (number of published articles is a very important metric in the academic world), to acquire new funds ("give us money to do research on topic X. Look at how much we published on this related topic before, we clearly know what we're doing") or advance in their career (ASA function of the former two points). Students need publications as part of their graduation requirements (in a lot of fields, you are required to have N publications in a peer reviewed venue to get your PhD).

So in both cases, you don't make money directly from publications, but from the prestige they bring and the credibility they're associated with ("this guy must be an expert on X, because they published on the topic").

I don't know, I personally have a slight preference toward hg, the UI feels a bit nicer.

par excellent

It's "par excellence", and it feels weird in that sentence structure anyways. A better way to put it would've been "my university has the low level computing systems curriculum par excellence".

As a member of the research community: that's nonsense. Like already pointed out: academic groups (who by no means are dependent on big tech) would jump all over that. Mamba has been out long enough that you'd already see tons of papers at arxiv showing mamba dominating transformers in all sorts of applications. But that's not happening, despite the ton of hype. That doesn't mean that mamba is nonsense. Just that it isn't the immediate transformer killer. It remains to be seen if something comes from it, eventually.

This sounds to me like you're extrapolating your own experience and generalizing it to the whole population, when in truth you are likely to be an outlier: most kids do fine in school (and always have), and most home schooled people find it harder to contribute meaningfully to society (due to less exposure to socialization opportunities and poorer education).

There is no “consensus” in the scientific method.

But there is consensus on how to use the scientific method, and sometimes the evidence accumulated with that method mostly agree with a specific part of the hypothesis space, in which case the consensus is to call that part the consensus