subject4056

I expect lots of humans to not be able to make you laugh on command, so no it's probably not a good test

Among other reasons, if you turn temperature down to 0, llms stop working. Like they don't give natural language answers to natural language questions any more, they just halt immediately. Temperature gives the model wiggle room to emit something plausible sounding rather than clam up when presented an input that wasn't verbatim in the training data (such as the system prompt).

I don't think so. Primarily because if you can ask that question instead of just being dead, then it's not the fast takeoff.

On a less drastic note, if AI were autonomously cannibalizing the economy, you'd run into more things and go "huh I guess that's run by AI now" instead of "ah godammit why did they shove an LLM in this workflow".

I don't think adding 20 bytes to every response, which are never read in practice, will improve the energy efficiency of the internet. Not to be mean, there's just a lot of stuff in most responses that effectively nothing acts on.

It's not a silly worry and you're not alone. https://pauseai.info/ .

We're not it's competition in the same way chimpanzees aren't our competition. Some fraction of us are interested in their well-being for aesthetic reasons, but a lot of the time this fraction loses to a not particularly powerful faction in direct competition for territory. And if there is any serious conflict of interest, there is no contest and the chimps lose. If we get lucky some fraction of superintelligence will look on us the way we look at ground apes, but that's far from a given.

I don't think it's right that Airbnb solved short term rentals - outside of a few dozens prestige markets that they monitor with humans, it's really a race to the bottom. Reputation at scale remains unsolved, and so it's still a market for lemons.

You could get healthcare from someone who doesn't have to go $100ks in debt first, and then expects to make more than you do. If no such person with that profile exists, then yeah you're stuck paying the rate of the system that trained and employs them.

At the end of the day, much like housing cannot be both affordable and a good investment, healthcare cannot be both affordable and a lucrative career. Avoid going to an MD if you can get the same care from someone else.

Kanban operates on a known product with a known manufacturing process. Many software products are undefined even at time of public release, and evolve continuously. "Deciding what to build", while explicitly highlighted in the agile software manifesto, is the weak link.

Put another way, lean manufacturing improves metrics for the margianal unit of goods. No one, customer or dev, is interested in the margianal unit of software.

I would guess not. An important feature of compilers is that they are guaranteed to emit code with certain properties in response to specific inputs (memory safety guarantees, asymptotic performance, calling convention, etc.). If they don't do that, you can file a bug report.

You cannot file a bug report against an LLM that it produced an unexpected output, because there is no expected output; The core feature of an LLM is that neither you nor the LLM developer knows what it will output for a wide range of inputs. I think there are a wide range of applications for which LLMs core value proposition of "no-one knows a priori what this tool will emit" is disqualifying.

Audio and video are encoded, compressed, and transmitted differently because of how humans audially/visually decode them. We fare better dropping late video frames than degrading their quality, where the opposite is true for audio. As we transmit the two as separate, asynchronous signals, it's unsurprising that they are frequently out of sync.

https://makesunsets.com/

If you're concerned that you've made/found something dangerous, the most appropriate solution is to disclose it to people who can evaluate how dangerous it is, and work with them on next steps.

In your hypothetical, I would observe that Paul Christiano is close to being the person in the US government in charge of evaluating AI danger, and that he is currently involved with https://www.alignment.org/ who specifically consider issues of AI risk. I would humbly contact them and ask for guidance on how to proceed.

Be prepared to be dismissed as a crank. Remember that avoiding harm to others is more important than proving that you're right.

Not significantly. Humans are made up of matter already available at the earth's surface, so population increase alone effects neither the amount nor distribution of the earth's mass.

Technological civilization might at some point meaningfully shift the distribution of mass, but I don't think it has up to this point.

You ought to begin by getting really good citations on each of those figures. If you or the source you got them from had any confusion about how to arrive at those numbers, any results you get from them will likely be meaningless.

Hopefully in doing so you will start to notice the missing pieces of info and can ask more targeted questions to get better results.

In order for passive investing to produce returns, stocks need to be regularly priced. Active trading produces the prices.

Active traders are trying to beat one-another, but at an institutional level the point isn't to beat the market, but to make the market.

https://simonsarris.substack.com/p/designing-a-new-old-home-...

Is highly opinionated but on-topic.

People are more impressed by things they cannot do than by things they can. The vast majority of people in the industrialized world are functional writers; the portion who are competent performers is much lower.

While professional writers may be quite skilled, the gap between what they do and what the median adult does seems traversal. The psychic distance is much larger for other creative endeavors.

Focus on something you expect AI not to exceed professional humans at in the next couple of decades. If you're convinced there's no such thing, then I wouldn't worry so much about career choice.

For a concrete answer in tech, have a look at AI development itself. Keep abreast of what OpenAI and their competitors identify as open problems and orient your academic career towards working in them. The object level questions will change while you're in school; the goal is an "intercept trajectory" where what you study just before you graduate is the state of the art.

Alternatively, highly regulated professions (medicine, law, civil engineering to name a few) are likely to continue to employ bright humans long after AI can do the job just because no-one will be allowed to use AI there.

A script is a type of program, they're not two separate classes of things.

I think classically a script is a program whose subunits are each complete programs ('commands') that operate on a shared environment, with shell scripting being the central example. The definition has since been expanded by way of analogy to include basically any interpreted program which may have side effects on its execution environment.

I tried a language with which I am unfamiliar, and do not think I could learn anything this way.

It was fun to practice with a language I'm functional with, but not a conversation.

I asked the thing if it was GPT-3 and it changed the subject to a very eliza-like prompt.

What's worked well for us is to adopt the framing that both of our time is equally morally valuable, if not equally financially valuable. One hour of my life spend doing work for pay is worth one our of her life spent doing work for pay.

We've operationalized this as: 1. Determine how much we want/need to spend on joint things (groceries, mortgage, vacation fund, etc)., call this B for budget. 2. Look at how much we're each making, M for me and H for her. 3. Find the fraction of our money X (which is roughly speaking equivalent to fraction of hours worked in our case) Such that XM + XH = B 4. We each contribute our respective portion to a joint account, and pay for all joint things out of the joint account. 5. We each keep the rest of our money in our own accounts, to do with whatever we like.

You might need to add a couple of terms to account for different work schedules.

Knowing that we're both contributing equal _effort_ to everything makes the inequalities (I contribute more to shared stuff, but also have a bigger personal discretionary budget) not seem like such an issue.

Beware survivor bias here. No-one regrets what they did well, but misses what they gave up to do well.

A very viable route is to spend the next 5 years maximizing your money and career. At 28 it is unlikely you'll regret options. (On that note, if you come from money, taking risks is the way to maximize your options).

I don't know if you have people (to talk to, rely upon, etc.). Most humans need to have people, if you don't have them then you should find some. In my experience this requires low level engagement over a long period of time, so there's no reason not to start early.

Tried to load game in firefox, rejected saying I need to run in desktop Chrome.

I was promised cyberpunk but what I got was a technologically mediated entertainment experience I couldn't participate in because I lacked the required proprietary software.

Was not dissapointed.

It sounds like your team's culture is built around a small dev team with a high level of autonomy and trust, which hasn't scaled to a larger org. Now's the time to establish a new culture.

Unless you are Milton from Office Space, you're not the only one on your team blocked by process (If you are, Amazon has great deals on bulk accelerants). Put out some feelers for others in the same situation. Junior team members are fertile waters. Schedule a twice-weekly recurring meeting to review one another's contributions. With discipline you can review two twenty line changes in an hour. Homework is unlikely to work here, or you wouldn't be posting this. In short order you will not only be pushing changes, but establishing yourself as someone who Gets Stuff Done.

Above all, don't sit at your desk and weep bitterly that nothing is merged. Stand up for doing things; either your team will improve, or you won't need someone else to tell you it's time to move on.

It's a little hard to tell from your description, but I will write a response assuming you're interested in studying PL theory as a career, and not just as an undergraduate. If you're interested in studying this area only as part of an undergraduate education, it really doesn't matter where you go since the content of your studies will be unrelated to your career, and you can study w/e you want on your own with very little impact on your future.

If you want a career in academia, DON'T pursue a bachelors from an institution that is at the top of the field you're aiming for. The most important thing you will do during your undergrad is making connections. This can be done more easily at middle of the road schools (where there's less competition and if you look good enough on paper better often better financial support) than at larger, more prestigious ones.

First, you will need close relationships with one or a few professors in the field who can advise (and recommend you) in applying for grad schools; these connections are critical. A useful approach can be to look for conferences/journals where "the professors [you] look up to" present their work and identify other authors (or coauthors) from less-well-known institutions, get in touch with them, and apply to the relevant programs where they work. If that doesn't pan out, go somewhere that (A) you've heard of (B) has a CS or good math program and (C) you beat the average admit stats. A big fish (like, class rank < 10th) in a small pond with good references has a lot better shot at being picked up than an anonymous 80th percentile sea bass.

Once you're in school, do good work, but your job is still to make connections. Cultivate relationships with the relevant professors in your program. Preferably have an advisor/mentor with whom you meet at least couple of times a month. Attend conferences, and once you've got a couple of years of study under your belt see if you can present at them (in many fields there are undergraduate conferences just for this purpose. probably no one with a name you recognize will attend these, but it's good to show you can defend your work to a sophisticated audience). As you get closer to applying to grad school you will want to identify potential advisors (more important than the school), and get in touch with them. It sounds like you've already got some people in mind, but it would be good to develop an understanding of the current state of the field and who's actively moving it forward (former students of Big Name professors are a good source of candidates).

As you observed, it's hard to convey to laypeople your interests and how your work relates to them. It is often not much easier with people who know what you're talking about. Good work is necessary but not sufficient; passion probably isn't even necessary but might help you mitigate the grind. To succeed you need to play the game better than everyone who applies to grad school but doesn't end up with a tenured position.

Alternatively, eschew college, go to lambda school or something and a year from now get paid six figures to sling javascript. Not great advice, but neither is grad school.

Fines like these don't tend to cause immediate personnel nor budgetary repercussions for the business unit in question, for the same reason BUs aren't punished if they miss sales targets for a quarter. Every BU's strategy is reviewed and signed off on by that BU's legal team before anything comes close to being shipped. A negative judgement isn't indicative of a mistake, it's suffering the loss of a calculated risk. If you do some VW level shenanigans you may get fed to the wolves, but most of us are smatter than that.

Ultimately the effect on the organization/employees is the same as any other financial loss. If it's too big or comes too soon in sequence with other losses, there will be strategic realignment (tm). Otherwise it's baked into the operating budget.

Source: Better part of a decade with Qualcomm

This doesn't appear to recognize certain common latinizations. For example, splitting ö int oe, as in "Foetzen". If you intend to commercialize this, you might want to invest some effort into that.

Stanford, MIT, CMU, UIUC, UC Berkely

UCLA, USC, and UC San Diego all make the list, but those are to some extent peculiar to my experiences in California. The last few are probably a different grab bag depending where in the country you want to end up.

I can't speak properly to how things are done abroad, but my understanding from other fields is that name recognition matters even more.