broast

I like how it's now a feature to not be an electron or webview app.

Ask Jim Simons

This doesn't add up to me. I thought "People over processes" was a pillar in the agile manifesto.

I've averaging a few conversations a day. My usage does not involve asking it to generate large chunks of code or anything mind blowing, but the usage I do get out of it makes me feel very fast, efficient, and confident in my abilities to tackle every little task I face every day that has any bit of unknown to me.

I don't get where this sentiment comes from. I build software specifically on the concept of predictable results from llm's being composable.

Sure, the results are not deterministic in that 100% of the time the exact prompt returns the exact same result, but you can tune your prompts so that 100% of the time they give you a valid result in the result category you were seeking, and with a specific probability distribution of available choices.

Prompts are functions that can take concrete input and create a probabilistic output that can be automated upon. Especially if you only need to output one token, i.e a number, boolean, word, object reference. And for obvious reasons - the further you forecast out in a sequence the less accurate you will be.

As long as you don't change the underlying model, in a massive model with billions of parameters, there are definitely mechanisms and behaviors to discover that you can reason about.

The post suggests it is 40% which seems correct to me.

Prompts are programs. They take input and produce an expected output. You can parameterize them and they can be used as functions and chained together. In that way, prompt engineering is software engineering.

I find it hard to believe that the intelligence you assume is required to do a person's job is an accurate indication of that person's intelligence.

Our remote teams have this culture that you say is impossible. They're in their chats or voice calls all day communicating, and measurably crush the in-office teams' productivity, maybe due to the efficiencies gained by going async, or maybe due to the type of people (people who find interactions through text more meaningful than in person or verbal, and insist upon it)

The in office teams on the other hand seem to rely on synchronous work and so they block each other, and this seems to mean the group can only finish a small number of tasks per day, whereas our remote teams complete bigger work more frequently.

I also find a lot of our juniors or quieter folk are a lot more confident talking to people in the remote setting.

They get more completed work out of the engineers they "over hired" before they have to let them go. Then they'll move 3 cheap engineers to maintain that shiny new product that 30 laid off engineers built.

I thought RLHF with LoRA is precisely the alignment method.

Since they mentioned processed foods, I think this could be mainly referring to ultra processed foods, where there is now a plethora of evidence that such foods contribute to poor health outcomes.

Rather than copying its output for publishing, I assume most people are using it to have conversations with, and ask questions, probe the answers, discover new information, research or use that information, and follow up with the conversation over time. That's how I use it and it's now my first step before Google, for certain types of questions or thoughts. I don't see that utility really going away.

"Intelligence, uh, finds a way."

You're definitely right that the news loves to write headlines after any daily price fluctuations and pretend they know the reason, when in reality it usually explains nothing and becomes irrelevant by tomorrow.

Our cells built our bodies and brain as an AI to navigate the macro realm!

Your experience matches mine other than I still find it extremely useful regardless of errors

I've been doing every-other-day-vegan for over a year. It was a good compromise for me, as a once meat obsessed person. It's actually a great system, I never eat the same thing two days in a row!

Sorry for any implied snark. I don't think it's a special set of skills, just a different default and preferred mode of experiencing human connection that some of us ended up with.

This actually has nothing to do with public forums. That was your insertion. The point is that the way we interacted with public or private niche communities on the internet informs how we interact with our private always-online work-chat. We're learning from and working with highly qualified coworkers aligned on the same problems, not randos.

Fair enough

I'll also ask that you consider the possibility that you haven't experienced overcoming this limit yet. After all, what do we need to learn about our work that can't be represented as text

Or that this limit might be qualitatively different for different people, and that some people are better at online text-only relationships than face-to-face ones. Our teams may have faced limits in real life that only got solved by going remote. (For example, I do believe that when we were in person, we wrote lower quality software, less well-tested, with less communication and collaboration, at a much slower pace, than we do now (you can imagine why - likely we used to rely on informal processes more than we thought or wanted to))

I did mention that I've seen other teams (anecdotally, on less technical areas of the product) face problems with being fully remote- but that was usually fixed with staffing... or a remote leader inheriting the team... We've had many people in our org exclusively remote for the last 20 years as well - these people became the remote leaders on our teams and led by example. It all depends on the remote fit.

Which some of us are way better at making online text-only relationships than face-to-face ones, which is how my career accelerated, and where most of my closest and longest lasting relationships in life have come from.

Well, I am sorry for that.

But just to add - it's a misconception to see remote work as a "new way" of working for someone like me. For someone like me, who used chatrooms and forums to collaborate with people on programming and similar projects, remote work is returning back to my natural way of doing things. Going to the office for 10 years was unnatural. We have now cut out a lot of overhead of learning how to manage office life, which is only relevant if you have an office, and which was a hard part of my life (I'm ugly).

And my experience is simply the opposite of yours. My strengths and my teammates strengths have only come to shine since going remote. I only became a Team Lead and then Engineering Manager since we've gone fully remote, and I think it is largely in part thanks to the speed and effectiveness my remote teams achieve in on-boarding members, collaborating, and delivering valuable work. Every retro my team members put "Great teamwork!" on the board, and this only started since we've gone remote.

I personally have experienced it, but my team has only seen greater success since going remote, and filling our roles with good remote fits.

I might say if you haven't experienced carrying a meaningful text-only relationship over years or haven't treated it equal to your "real life" relationships, then you can not know the effectiveness that comes from it

Your take is very dismissive and does not align with my experience. We work on complicated technical high value things and my juniors are very fast learners. And we learn from each other and not necessarily from the internet.

We also don't need video calls. I have never met them, and rarely have I seen their faces. Kind of like my life long friends on the internet, we didn't have video calls then either. We can connect through text just fine. We sometimes play games together to blow off the steam, and we collaborate on meaningful work problems.

Personally, I grew up on the internet, learning from people in chatrooms and forums which were strong communities where most of my personal adult relationships have come from (many of whom I've not met in 20 years of knowing them).

From what I've seen, if your seniors and your juniors are both from this era and culture, your remote team will excel without missing a beat. I think it's the people used to the "old ways" that will suffer, for having expensive requirements like an office to capture the same value.

Prompt engineering is just communication skills. I could tell you which of my coworkers would make good or bad prompt engineers just by looking at their transfer notes on tickets.

An important distinction is while fasting for Ramadan, you can't ingest anything at all, including water. Unlike intermittent fasting where practitioners might still have 0 calorie coffee to stay stimulated, and water to stay hydrated.

Just Jim Simon's own words, in the book The Man Who Solved The Market. And the fact that they hired NLP engineers explicitly.

My guess is it would be momentum based and capturing short term fluctuations in the market. My guess is they would train a markov decision policy on sequences of moves, and try to predict the next one, and trade on it with some sophisticated risk management.

It's almost certainly expected.

These models are argued to contain hidden models of the data they represent, and can pick up on patterns with one-shot or few-shot samples

Let's recall that Bob Mercer, Jim Simons made their billions off of trading options with NLP based hidden markov models, beating everyone on Wall St for the last two decades.

ChatGPT is an easy entry to this