Feb 3
The Most Interesting Tech Reads in January 2023
mostly fascinating takes on AI, but I’m also cheating and sneaking in some politics!
The aim of this page📝 is to share my monthly bookmarks. Book-wise, I’ve been reading the famous Pragmatic Programmer by Dave Thomas and Andy Hunt and Black Earth by Timothy Snyder
GitHub Copilot preliminary experience report
- https://blog.ploeh.dk/2022/12/05/github-copilot-preliminary-experience-report/
- Notes on Experience with Github Pilot, similar to that mine — it has the potential to surprise you, which is great.
I was, frankly, stunned. While I do understand that Copilot doesn’t understand the code it suggests, Hedgehog is sufficiently esoteric that I didn’t expect Copilot to have enough training data to enable it to make a useful contribution in this niche. I was wrong. I’ve repeatedly seen Copilot make suggestions to my F# and Haskell code. It’s not just for C#, JavaScript, or python code.
Free YouTube Transcripts
- https://ytscribe.com/
- For cases a transcription’s required
imaginAIry/README.md at master · brycedrennan/imaginAIry
Overconfidence and AI
- https://seths.blog/2023/01/overconfidence-and-ai/
- My friend told me “AI is bullshitting a lot” which Seth Godin kindly frames as overconfidence
Human beings are often more effective when we’re a bit self-effacing. “I think,” “Perhaps,” or “I might be missing something, but…” are fine ways to give our assertions a chance to be considered.
Fighting Distraction With Unit Tests
- https://matthewc.dev/musings/unit-tests/
- A Hemmingway-style approach to writing code where you stop at writing the failing unit test as a sort of cliffhanger.
ChatGPT Is Only Going to Get Better, and We Better Get Used to It — Bloomberg
- https://www.bloomberg.com/opinion/articles/2023-01-23/chatgpt-is-only-going-to-get-better-and-we-better-get-used-to-it?utm_content=view&utm_source=twitter&utm_medium=social&utm_campaign=socialflow-organic&cmpid%3D=socialflow-twitter-view&sref=htOHjx5Y
- AI is developing in leaps — mind you.
I’ve started dividing the people I know into three camps: those who are not yet aware of LLMs; those who complain about their current LLMs; and those who have some inkling of the startling future before us. The intriguing thing about LLMs is that they do not follow smooth, continuous rules of development. Rather they are like a larva due to sprout into a butterfly. It is only human, if I may use that word, to be anxious about this future. But we should also be ready for it.
How to Profile Your Code in Python
- https://towardsdatascience.com/how-to-profile-your-code-in-python-e70c834fad89
- I am wondering why a function that I expect to take shorter is taking longer for leet code puzzles
Premature optimization is the root of all evil.
ChatGPT’s killer enterprise use case will be managing knowledge, says EY CTO
- https://venturebeat.com/ai/chatgpts-killer-enterprise-use-case-will-be-managing-knowledge-says-ey-cto/
- Knowledge keeping, managing, and providing seems to be killer feature for Enterprise AI
The coming ubiquity — Seth’s Blog
- https://seths.blog/2023/01/the-coming-ubiquity/
- On the invisible revolution of making AI everywhere all the time
But the real impact of AI isn’t going to be that it regularly and consistently does far better than the best human effort. The impact will be that it is widespread, cheap, and always there.
SQL should be your default choice for data engineering pipelines
- https://www.robinlinacre.com/recommend_sql/
- After a wave of NoSQL tools, this is what you get in the beginning of 2023
SQL should be the first option considered for new data engineering work. It’s robust, fast, future-proof, and testable. With a bit of care, it’s clear and readable. A new SQL engine — DuckDB — makes SQL competitive with other high performance data frame libraries, making SQL a good candidate for data of all sizes.
dylanjcastillo/shell-genie: Your wishes are my commands
- https://github.com/dylanjcastillo/shell-genie
- Another cool project to get info on AI
Shell Genie is a command-line tool that lets you interact with the terminal in plain English. You ask the genie what you want to do and it will give you the command you need.
Natural language is the lazy user interface — Austin Z. Henley
- https://austinhenley.com/blog/naturallanguageui.html
- A slight critique of chatgpt as an interface as it lacks a big advantages of adaptable GUI.
People are anticipating that large language models are going to revolutionize the world. And maybe they will. But a chatbot won’t.
On Email and Horses — Study Hacks — Cal Newport
- https://www.calnewport.com/blog/2023/01/25/on-email-and-horses/
- Cal Newport’s ideas about smartphones being dangerous for anyone under 15 are shocking
An iPhone is a superior device to a Motorola Razr, but does this mean 12-year-olds should be using them?
The inside story of ChatGPT: How OpenAI founder Sam Altman built the world’s hottest technology with billions from Microsoft
- https://archive.ph/2023.01.25-131857/https://fortune.com/longform/chatgpt-openai-sam-altman-microsoft/
- History of OpenAI
Oleksiy Danilov interview:
The West has an old problem: that is fear. The West has always been afraid of the Soviet Union. It believed the USSR was big and powerful and could solve everything quickly through military means. No conclusions were made even after the USSR invaded Afghanistan in 1979. And they couldn’t do anything there against the uneducated, at that moment, the mujahideen. That is when a conclusion should have been made that the USSR is not that powerful.But those conclusions were not made. And the West kept on being afraid.
Five Days in Class with ChatGPT — The Alperovitch Institute
And I say this as somebody who had been a hardened skeptic of the artificial intelligence hype for many years. Note that I didn’t say likely transform. It will transform higher education. Here’s why. The first use-case is that the machine “filters mundane questions,” as one of our students put it quite eloquently. Means: you can ask dumb questions to the AI, instead of in-class. Yes, there are dumb questions — or at least there are questions where the answer is completely obvious to anybody who knows even just a little bit about, say, malware analysis (or has done their assigned readings).
Get/Want/Have To
- https://seths.blog/2023/01/get-want-have-to/
- A poetic approach to a mindset — for a thief of moments for passionate endeavors, like reading a book for 5 minutes.
The magic trick begins with realizing that the get-to tasks are priceless want to moments if we choose. And, if we’re careful and plan ahead, we can get to the point where the have to agenda is something we can eagerly look forward to.
Alex Epstein’s Fossil Future — Marginal REVOLUTION
- https://marginalrevolution.com/marginalrevolution/2023/01/alex-epsteins-fossil-future.html
- Remember Macaes: Most dangerous political movement in the West today is not populism, post-fascism, wokeism… It’s a reflexive contrarianism that affirms the opposite of every sensible opinion. Silicon Valley actually played a role in this. Idolatry of disruption but now out of its element.
My overall view is this: it is a good rebuttal to “the unrealistic ones,” who don’t see the benefits of fossil fuels. But it does not rebut a properly steelmanned case for a transition away from fossil fuels.I view the steelmanned case as this: we cannot simply keep on producing increasing amounts of carbon emissions for centuries on end. We thus need some trajectory where — at a pace we can debate — carbon emissions end up declining. I’ve stressed on MR many times that climate change is not in fact an existential risk, but it could be a civilization-destroying risk if we just keep on boosting carbon emissions without end. I don’t know a serious scientist who takes issue with that claim.
Yiren Lu on Twitter: I’ve been playing around with GPT for various use cases. Here are some low-hanging ones I haven’t heard people mention
- https://twitter.com/YirenLu/status/1620171019788120066
- Continuing the discovery of use cases and cultivating the general awareness of the arrival of the beast.
At my agency, a great deal of operational work involves simply taking information from email conversations and inputting them in structured ways into a spreadsheet or web app. This is quite tedious to do and is prone to errors/missing data. But ChatGPT is pretty good at this!
