Vivek Haldar

Some thoughts on the anti-LLM sentiment in the programming community, sparked by this article.

The AI as a product vs as a feature lens of looking at AI is going mainstream, with Humane and Rabbit exemplifying “AI as a product” and Apple Intelligence representing “AI as a feature”.

Practitioners know that there is no such thing as prompt portability right now.

Programming with LLMs, and building systems with LLMs, has completed the long de-nerdification of the field.

(This is an edited write-up of my rambling video on the same topic.)

Ever wished you could provide ChatGPT one initial prompt, then kick back for a few minutes as it waxes eloquent on that topic?

Not so long ago, I suggested that when it comes to using LLMs for programming, a useful metaphor is to think of them as compilers.

Ed Summers wrote a post on why he doesn’t use Copilot, and I want to riff on that to see if it helps me build my own mental scaffolding for how LLMs, and AI in general, will change programming, and how programmers should adapt to that.

(The following is a slightly edited script of my video on the topic) The AI doomers are everywhere, and they want to start blowing up stuff, and I want to talk about it.

I was re-reading Charles Petzold’s 2005 talk/essay Does Visual Studio Rot the Mind?

I’m confused and in a funk, hence this lazypost.

I’ve been thinking about a fruitful way to frame the act of writing code in the age of Copilot/Codex, and I don’t think “autocomplete on steriods” is it.

I’ve been writing Python since version 1.5.2, and I think Guido von Rossum is one of the best language designers of our time, but I was a bit taken aback when on a recent Lex Fridman podcast he said, “To me, VS Code, in a sense, is a spiritual successor to

Libraries like LangChain show how to use the output of LLMs to drive actions through external tools and APIs.

It is now well-known that arithmetic is a weak spot for LLMs.

Took a bit of elbow grease with miscellaneous Emacs macros and tiny bits of elisp, but I’d been meaning to move my writing off Tumblr and on to a properly self-hosted place.

Here’s a collection of some hardware acquisition hacks I’ve been using over the last few years.

Over on my YouTube channel, I make talking head videos, covering various computer science topics, mostly research papers.

Just thought I’d put this tip out in the Internet… So I recently got a Thinkpad Trackpoint II keyboard, which I like to use with my MacBook when docked to an external monitor.

Unix has won in every conceivable way.

I’ve been putting up a video a week on my YouTube channel since about December 2018, and I suppose I should write out a clear manifesto for why I’m doing that and where I want to go with it.

I love a good keyboard.

Chaos Monkeys: Obscene Fortune and Random Failure in Silicon, by Antonio Garcia Martinez People outside Silicon Valley may think of this book as entertaining and illuminating, but if you’re part of it (even if not geographically but within it’s extended m…

Some scattered thoughts after reading about the airspace aesthetic.

Just because it’s fun to play with the natural numbers in wildly impractical ways.

Via Marginal Revolution, I came across this paper on a phenomena known as algorithm aversion: … evidence-based algorithms more accurately predict the future than do human forecasters.

I’m making my way through Nicholas Carr’s latest book, The Glass Cage–slowly, because life.

Nicholas Carr’s new book, The Glass Cage, opens with a discussion of the FAA’s recent 1-page alert that cautions pilots to do more manual flying.

I have often wondered how things grow, and specifically, how software projects grow.

Spolsky’s famous screed against rewriting from scratch is now canonical.

If you are a “regular” person living and working in modern civilization, it is likely that your life has a few recurring patterns: on weekdays, you travel from home to work in the morning, and back in the evening; perhaps punctuated with a school or dayca…

I define a garland word to be one which can be formed by chopping off the last few letters and “wrapping around” to the start.

Stephen Kell’s paper at PLOS 2013 meditates on one of CS’ oldest design questions: what should be in the operating system, and what in the programming language and runtime?

It’s commencement–and commencement speech–season.

Tech’s secessionist streak has recently got a lot of attention.

A lot has been written about Heartbleed.

I recently read “The Second Machine Age” by Erik Brynjolfsson and Andrew McAfee.

Ex-Googler Amit Patel made an insightful comment on my post about fat and thin clients that I totally agree with, and so I’m just going to reproduce it here in full: Some thoughts: 1 - Net connectivity does not necessarily mean “thin client”.

Maha Bala and Brad Meier take on the default assumption that synchronous modes are best for online learning.

Corporations can be thought of as information-processing feedback loops.

I’m growing increasingly frustrated with how little today’s patchwork of messaging/email and docs matches my workflow needs.

I’ve been following Venkatesh Rao’s writing since he achieved net stardom with his Gervais Principle essay.

It used to be that the platform on which most software was developed was also the one on which most of it was deployed—the personal desktop computer.

If you look at features in mainstream programming languages over the last couple of decades, they seem to be converging around a common set of core features, surrounded by fashion, preference and whim.

I’ve followed Kevin Kelly’s work for a while, and in my humble opinion he is our foremost apologist for technology.

Many years ago, before I had a GPS-enabled smartphone with maps, we went on a roadtrip along the coast of Maine, stopping at nearly a dozen lighthouses.

Star Trek and Babylon 5 are two space operas that offer very different views of humans, technology and the relationship between them.

Ben Thompson’s latest post is about Dropbox, Box and focusing on consumers vs enterprise as customers.

Lately I’ve felt lost when in front of the computer.

So I finally got to use Go at work, which was a great excuse to learn the language.