diego

Meta has extremely opaque account policies. For example, I bought the Meta Raybans a month ago. It kept telling me the AI features were not available in my region, even though I am in San Francisco. I joined Facebook in 2006, and I have used my account for the Oculus headset without a problem. But no matter what I did, the AI function of the Raybans wouldn't work.

I ended up creating a brand new account just for that, and it worked fine. No idea why it would work with a brand new account and not with my old account in good standing, never suspended or warned about anything.

I think the title just resonates. Older folks like me must reflexively upvote this stuff. I admit I do sometimes.

correct link, @dang please fix https://iamnotarobot.substack.com/p/why-i-am-starting-a-hard...

https://iamnotarobot.substack.com/

Lately I've been writing about AI because it's impossible not to. But I generally write shower thoughts about tech and my experiences in the industry.

No complaint. It's more of a warning about how the main players (OpenAI, LangChain) share notebooks and cookbooks that illustrate how to make the LLMs "query" the databases. At the very least one would expect some language telling people to not do that in production. And it's not unique to SQL, this is just an extreme example.

tl;dr: nothing we didn't know. Since the beginning of times, startups with lots of funding have failed for a number of reasons. AI is no different in that regard.

Seconded. Start here, and pause whenever you realize there is something that you need to learn in order to follow. Learn that and keep going.

I had not seen that article. I'm not surprised someone had a similar idea before.

If you try Bard or Claude or character.ai they are not far behind GPT4. They might even be on par in terms of raw LLM capabilities. ChatGPT has better marketing and in some cases better UX. A lot of this is self-fulfilling. We think it's far ahead, so it appears to be far ahead.

That's exactly what I did here. https://github.com/dbasch/semantic-search-tweets

Apparently it's simply a writer's website.

https://en.wikipedia.org/wiki/Lucy_Ives

I mostly agree. The way I would face this decision is like this:

1) Do I enjoy the coffee business? Would I have bought that plantation if the opportunity had arisen? If the answer is no, selling is most likely the right answer.

2) That being said, it's not like the plantation is a hot potato that needs to be sold right away. OP has some time to check it out, learn about the business and decide if this is something worth trying for a few years. How long to try for? What's the opportunity cost? They would decide to invest a limited amount of time (e.g. one year of work) and then reevaluate at the end. I would timebox this decision process too. Perhaps learn all you can in a month and decide if you want to give it a shot? Part of this process should involve getting in touch with people in the industry and seeing if you like interacting with them. In the end, most of your job will be about interacting with the different types of players in your industry.

Large companies have lots of employees not because they need them, but because they can. Many of the people who are good at leadership happen to like empire building and having reports. Every employee in a company is also an evangelist and a marketer, without even trying. Everyone answers the question "where do you work." Having lots of employees gives you slack, and it's easy to get rid of the slack with layoffs.

It must fit in a similar space to a human. Like it or not, it’s what the world is designed for.

This doesn't follow. The machine can be distributed, only peripherals need to fit in human space and they don't need to be smart.

Discipline is what you have to use when you don't have a habit. Once you have built the habit to do the thing, discipline doesn't enter the picture. It's the default. You don't buy candy because buying candy is not a thing you do. You go running because it's 9 am on a Monday and that's what you do every Monday at 9 am.

I just tried Copilot with VS Code and python for the first time. If I define a function with some parameter name, I get suggestions as I type the body. I change the parameter name to gender, no suggestions. I change one letter in the parameter name (gendes, gander), I get suggestions again. There clearly is some code that gets activated when it sees the word "gender".

From the article:

The Royal Mail is aiming to deploy a fleet of 500 drones to deliver mail to remote areas, starting with small islands like the Shetlands or Orkneys. Britain’s National Health Service is trialing drone delivery for medication, specifically chemotherapy drugs, to cancer patients living on remote islands; one such route would save patients a three to four-hour ferry ride to the mainland and back to pick up their medications.

Obvious business idea: toilets that slide a containing partition before flushing. There's no need to use the seat for that, or to require the person to get up.

There might be easier mechanical ways to accomplish the same.

There are certain kinds of problems that don't allow you to overthink. If you are on a tight deadline and you need to get something done (e.g. fix a bug, add a feature) then the situation forces you to make _a_ decision and move on. You can only overthink when you don't have hard deadlines. So as an exercise, try to pick challenges that have deadlines (ideally external to you). It helps to work with other people, take up on some responsibility and commit to delivering.

Nobody needs to travel with 8 pairs of underwear. Take two or three high-quality ones (e.g. Ex Officio) and wash them at night.

Also, I have yet to visit a civilized location in the world where you cannot buy a decent disposable razor that's better than the electric shaver he carries around the world. I understand they recycle pretty well too, if you're worried about waste.

The reason people keep using Twitter it's because it's designed / evolved to be very addictive. Most of us don't engage our rational mechanisms to go "I am about to spend X minutes engaged on Twitter, what's the cost/benefit?" We just do it, and later feel bad about it.

If you've been on Twitter since the early days, you have noticed how it has become more and more about dunking and hate-quoting. Once you have an algorithm that decides what content causes the most engagement, you learn that negative emotions make people act. It's hard to scroll Twitter and not want to yell at something you disagree with. The medium itself does not lead to constructive conversation. This comment would require a Twitter thread, which is a scary thing to write. Every single tweet in a thread can be taken out of context, and you have to write accordingly.

As a company, Twitter has stopped innovating a while ago. The product has been stagnant, and so has the market cap. I expect that eventually it will collapse when the most interesting people to follow simply stop engaging, but that might take years.

My solution is to engage as little as I am capable of. I deleted the app from my phone, and I have a personal rule of only posting inoffensive shower thoughts that almost never get any reactions. I never get involved in an argument there.

Probably because many people who follow him on Twitter have no idea what running a startup is like, so they are imagining a completely different context. Business plans do make sense in other contexts.

Agree with that. However, in 2022 writing is not the only way to order your thoughts outside your head. I for example have put together Youtube videos without writing anything down. The process is like this:

- Think of a point I'd want to make, perhaps in 20-60 seconds.

- Do a few video takes until I believe it's good enough. Sometimes the first take works.

- Keep doing it until I have enough material, then edit the video.

As someone who is in his 50s though, I can understand the point of view of someone who grew up in an age in which writing was the obvious way. Paul might think differently if he tried to start a Youtube channel, but it doesn't seem to be interesting to him.

And someone who never writes has no fully formed ideas about anything nontrivial.

This would imply that Socrates only had trivial ideas. Socrates would disagree with this statement, and so would a large number of people. This belief seems like a cognitive bias to me. If you prefer written ideas, unwritten ideas may seem trivial to you.

No, but if you don't have your half memorized and they take it from you, the other half is useless. This is more useful if you want to leave your crypto to your family if you die, provided that you make it easy for them to find your half if you're not around.

An analogy is shorthand for: here is something relatively complicated. We can look at it with the same model that we have agreed upon for this other thing.

If the two things are similar enough in the ways that the model applies, then it's good. Otherwise, it's not.

Think of all the technologies that we use today that were invented at companies that no longer exist, or are no longer relevant. E.g. Java was invented at Sun, C at Bell Labs. Why would it be different?

That is an opinion. The 2001 crash caused a crisis, then a recovery, then two decades went by with many ups and downs. The entire peg might have never happened and Argentina might still be where it is today. Are you old enough to remember the hyperinflation of the late Alfonsin period? I lived through it. No peg preceded it.

Exactly. The reason inflation happens is that the government overspends, then prints money because it can. There are no magic tricks to stop inflation if you're constantly devaluing your currency.

Study Argentine history. They have tried everything imaginable, and managed to control inflation for a while. Look at the 1-1 peg between the peso and the dollar that lasted 9 years. It's much more complicated than you think.