taylorlapeyre

Well, LLMs have an obscene amount of context built into their weights about Ruby on Rails, and can work within it extremely quickly.

The infrastructure piece is what they do best. I'd be happy if Anthropic became the AWS of AI. But this product is just a mediocre shot at Figma, when no such shot is strictly necessary for them. These kinds of consumer products are not what they do best.

If that is what they are optimizing for, then they are not doing a great job on that either. Their latest model is more expensive and less impressive, their downtime is abysmal by any comparison, their "Dispatch" product is unworkable - conversation threads don't sync properly, the UX is quite bad -, their Desktop app is slow and clunky, Cowork is by all accounts mediocre.

It really feels like Anthropic's product area is extremely overextended at this point. If they want to extend themselves horizontally in an unlimited fashion, they will need unlimited focus, and agents can't supply that. Things will fall through the cracks. Why should I believe that Anthropic will care about this product in 2, 3 years? Whereas I firmly believe that Figma will care greatly about its product in that time

Sex is not like other subjects, it is peculiarly liable to harm, obsess, distract us. Sexual content is uniquely powerful in its ability to reduce human beings to objects for use, and having the ability to generate endless amounts of it completely privately is a very dangerous power to have. I worry about what it will do to us to have this technology, and I worry that OpenAI is treating this subject with too little concern.

I've recently read the great moral philosopher Alisdair MacIntyre's book "After Virtue", and in it he defines a "practice" as:

> "any coherent and complex form of socially established cooperative human activity through which goods internal to that form of activity are realized in the course of trying to achieve those standards of excellence which are appropriate to, and partially definitive of, that form of activity, with the result that human powers to achieve excellence, and human conceptions of the ends and goods involved, are systematically extended. Tic-tac- toe is not an example of a practice in this sense, nor is throwing a football with skill; but the game of football is, and so is chess. Bricklaying is not a practice; architecture is. Planting turnips is not a practice; farming is. So are the enquiries of physics, chemistry and biology, and so is the work of the historian, and so are painting and music. In the ancient and medieval worlds the creation and sustaining of human communities-of households, cities, nations-is generally taken to be a practice in the sense in which I have defined it. Thus the range of practices is wide: arts, sciences, games, politics in the Aristotelian sense, the making and sustaining of family life, all fall under the concept."

Programming is a practice (especially during the golden era of open source software), with its own "internal goods" such as described by this article: the pleasure of optimizing an algorithm, the "ah-ha" of finding a great root cause, the beauty of a well-written function, the fun of it.

MacIntyre also says that practices can only be incubated and cultivated within "institutions" - organizations which specifically exist to protect the development of a practice from the intrusion of external goods, by careful management of external goods. But institutions can become corrupted and degrade the practices within them. And indeed recently programming has been degraded into a simple skill used to obtain external goods, namely wealth and fame, and the institutions where programming tends to be cultivated tend to have deeply corrupted themselves. One can still recognize people in tech companies that fight against this tendency, but it's a remarkable confirmation of his thesis in my opinion.

Wait — isn't that exactly what good investors do? They look for what stocks are going to beat expectations and invest in them. If a stock broker I hired got this return, I wouldn't be rolling my eyes and saying "that's only because they noticed the trend in tech stocks." That's exactly what I'm paying them to do.

The post mentions why - Bun eventually wanted to provide some sort of cloud-hosting saas product.

It once again completely fails on an extremely simple test: look at a screenshot of sheet music, and tell me what the notes are. Producing a MIDI file for it (unsurprisingly) was far beyond its capabilities.

https://chatgpt.com/share/68954c9e-2f70-8000-99b9-b4abd69d1a...

This is not anywhere remotely close to general intelligence.

yes, certainly

I agree with you. Is it perhaps because of his religious background (he was a Catholic priest)? For much of the last couple decades, there has been an anti-religious streak in the educational mainstream universities.

Any God, as long as it's not a being/creature. Thankfully there's only one of those! https://en.wikipedia.org/wiki/Classical_theism

I agree, but with a caveat about ice cream trucks specifically. If they park in one spot and play the same jingle very loudly for hours on end (say, three to four hours in one location outside of your home), it is in fact insanely maddening.

The deep-ocean vent south of Antarctica is real but small, on the order of a few-tenths Pg C yr⁻¹. The claim that it could double atmospheric CO₂ exaggerates the flux by three orders of magnitude relative to observed values and known physical limits.

The most optimistic estimate of deep-water outgassing south of 60 ° S is 0.36 Pg C yr⁻¹. Even if that rate tripled and persisted unabated, it would take more than 800 years to add 895 Pg C (which would be what it would require to justify the press release’s claims of “doubling”)

What the salinity reversal can do is:

- Expose ice shelves to warmer subsurface water, accelerating sea-level rise.

- Reduce the Southern Ocean’s role as a sink by a few tenths Pg C yr⁻¹, nudging the global ocean sink (~2.7 Pg C yr⁻¹) downward.

- Perturb atmospheric circulation patterns, with knock-on effects for the Atlantic overturning (but those links remain speculative).

Paul, I’m begging you to read the Gorgias!

People really avoid considering what the word "general" implies. Yesterday I tried sending o3 a screenshot of some sheet music, asking for a midi file of how it sounds. Complete failure x3. Could not even get the value of the first note right. This is not "general" intelligence.

Well deserved, in my opinion. I often think about the fact that they spent 5 years (2012-2016) building essentially a brand-new graphics editor entirely in the browser way before Web Assembly was a thing that people knew about, all before even releasing their product to the public. They took the performance of their product so seriously.

I think their team deserves credit for their commitment to good software. They got here by making something good. That's a success story that I feel good about

What is a "late interpretation" in your view? What time period exactly? Job has been read as a story that directly dialogues on the possibility of faith and trust in God in the face of evil for millennia. See The Babylonian Talmud (Bava Batra 15a-16b), Midrash Rabbah (Genesis Rabbah, Exodus Rabbah), Saadia Gaon's commentary on Job, even Maimonides. And on the Christian side, Take Origen, John Chrysostom, and especially Gregory The Great's Moralia in Job in the 500s CE.

A toddler dying of cancer falls under the category of "natural evil" (as opposed to moral evil, in which an intention or will is involved). You're correct to say that natural evil can sometimes be difficult to explain in a pedagogical sense. A storm taking one's home can be a way to teach us our dependence on God alone, but a child being taken by cancer does seem meaningless to us at the best of times, and simply cruel at the worst.

But the story of Job (in which his family, including his children, are taken away through natural evil) shows us two ways of engaging with this kind of evil: 1) Seeing the resulting struggle with God as its own form of growth, and 2) the virtue of humility.

In my opinion, there is no "one answer" to Theodicy - it requires all three and a healthy dose of Job's eventual humility. The first has been true in my life multiple times. The second is often helpful when engaging with evil on a cosmological level. The third is helpful when wrestling with the problem of moral evil. Taken together, I think the picture is reassuring, and certainly better than the alternative (a meaningless universe of hopeless suffering).

I recommend David Bentley Hart's "The Doors of the Sea" as a good short work on this subject!

Technically correct, but quite misleading. The idea of "trust in God" or "faithfulness" is completely central to Job. The story doesn't concern itself with "doctrinal faith", but it implicitly discusses "faith" in the general sense of trust in the providence of God in the face of challenges that might make one abandon Him.

This is one of three major answers to the problem of Theodicy (justifying God in the face of evil) in Christian theology. They are:

1. Pedagogical (the one you mentioned) - Evil exists so that we may grow and learn. All evil will be used to create even greater good. Quintessential example would be the story of Job in the Old Testament.

2. Eschatological — Evil is a by-product of having any creation in time whatsoever, and will be fully restored at the end of all things. Another way of thinking of this one is that evil is "non-being", the absence of the good. A lack, a privation. Espoused by Augustine (and before him, Neoplatonism in general), and Aquinas.

3. Freedom-oriented — Evil (even natural evil) is a broken state of affairs caused by the freedom of people, who use that freedom against God. God nonetheless allows this because he allows us the dignity to choose. The official teaching of the Catholic Church. The straightforward reading of Genesis 2.

None of these say that "This <bad thing> is not bad" - Christianity acknowledges the existence of evil "as evil". However, with God's grace, evil may be healed or made to serve higher purposes.

I wonder what would a ceremony like this look like before the era of cameras broadcasting every move to the entire planet simultaneously. It does seem like a lot of this is performative in the “showtime” sense, and I wonder if the character of an event like this would be more “practical” in another time.

Yes! I often put it like this: education/reading is about formation, not just information.

By default, Figma will not using organizations' data to train.

They are extremely nice compact SUVs (Jaguars), extremely nice interiors with displays (like Teslas) that show the entire surrounding area with visualizations of what the car can see, you can choose music and air conditioning, and the driving is extremely good — equivalent to a typical safe (albeit slow) Uber driver.

Living here for the last 10 years, it's been jarring how just in the last few years, driverless taxis went from "it'll never happen" to "is this the default now?"

The Waymos are genuinely good drivers. I look forward to taking them every time.

There is no contradiction - it is reasonable to feel more than we do.

Helen Toner funded OpenAI with $30M, which was enough to get a board seat at the time.

They aren't stars, they're galaxies!

Figma bought [Diagram](https://diagram.com) this year, so this is inevitable and unsurprising