mordymoop

I found it frustrating that the article highlights Mullenweg’s actions while failing to even briefly describe what WP Engine is, and what actions they took. The story felt very incomplete, perhaps intended for someone who is already familiar with all the details. I have copied below excerpts from the Wikipedia entry on WP Engine because I found it clarifying.

> WP Engine's main function is allowing businesses and organizations to build, host, and manage websites powered by WordPress.

> During the week preceding September 22, 2024, Matt Mullenweg—founder of WordPress.com—began speaking negatively about rival WP Engine. Mullenweg gave a speech at WordCamp US 2024 that argued that WP Engine had made meager contributions to WordPress compared to Automattic, criticized WP Engine's significant ties to private equity, and called for a boycott, sparking internet controversy.[30] In response, WP Engine issued a cease and desist against what it characterized as defamation and extortion, attributing his attacks to WP Engine's refusal to pay Automattic "a significant percentage of its gross revenues – tens of millions of dollars in fact – on an ongoing basis" for what it claimed were necessary trademark licensing fees (later clarified as 8% of all revenue, payable in gross or in salaries for its own employees working under WordPress.org's direction, combined with a clause that would've prohibited forking[31]) for the "WordPress" name.[32] Automattic responded by sending its own cease and desist the next day, citing the trademark issue.[33] On October 2, 2024, WP Engine sued Automattic and Mullenweg for extortion and abuse of power, which the defendants denied.[31] As a result of the dispute, WordPress.org blocked WP Engine and affiliates from accessing its servers—which include security updates, the plugin and theme repository, and more—on September 25, 2024, a day after its trademark policy was updated[34] to ask against usage of WP "in a way that confuses people", listing WP Engine as an example.[35] Following backlash, access to WordPress.org was temporarily restored until October 1 to allow WP Engine to build its own mirror sites two days later,[36][37] which the company did.[35] On the 12th, WordPress.org replaced the listing of WP Engine's Advanced Custom Fields (ACF) plugin on the WordPress.org plugin directory with a fork called "Secure Custom Fields" citing a guideline that empowers the foundation to "make changes to a plugin, without developer consent, in the interest of public safety".[38] On October 7, 2024, to align the company's stance, Mullenweg announced that 159 employees—8.4% of Automattic—had quit in exchange for a severance package of $30,000 or six months of salary, whichever is higher, with the condition that the resigned would not be able to return.[39] The next week concluded another offer of nine months' salary to attempt to placate those who could not quit for financial reasons,[40] though with only four hours to respond and the added term of being excluded from the WordPress.org community.[35]

We came within a literal inch of witnessing the assassination of a presidential candidate earlier this year, by a kid with no particular skills and an easily obtained rifle. We are lucky that people are mostly nonviolent.

They make me feel like becoming super rich is achievable — even they could do it!

Whenever I think about automation, I, of course, think about having toddlers in the house.

Toddlers are a transient disruption to the household logistics operation. They induce novel challenges. Their clothes are have different form factors, and may be dirtied in novel and interesting ways. Their nutritional demands are only partially-overlapping with the preexisting household diet. They introduce hundreds of small toylike objects into the environment. Even their dishware is smaller, flimsier and more numerous, disrupting any existing dish-washing workflows. And the development (or purchase) of permanent solutions to these problems is in some sense wasteful because the toddlers will stop being toddlers before too long, and you will have to pivot again.

The holy trinity of washer, dryer, dishwasher solves a relatively high percentage of household labor but the tail is very long. The purpose of the above digression on toddlers is to illustrate that some household labor disruptions are transient and really not worth developing specific solutions to permanently resolve.

This household example is both a true example of the problem and a metaphor for this type of situation which occurs in all human environments. There is always a long tail. There are always changes to processes with knock-on effects that leave gaps in preexisting workflows. And the solution of final resort has always been human manual labor. There are always transient solutions that require transient workarounds. And this long tail is where it will be enormously helpful to have humanoid form factor robots. This is true for exactly the same reasons that a young parent would appreciate having a humanoid robot in the house, to pick up clothes off the floor and fold clean ones, gather up toys and put them in their appropriate storage bin, and clean the table and floor and nearby walls (and possibly ceiling) after meals.

I have worked in specialist consulting for 10 years and I agree with this take. Our business does the best during times of uncertainty. During lean times, operators want to make sure they are allocating capital intelligently, so they pay for our insights and advice. In the good times, they just plow all their revenue into making the flywheel spin faster, no need to be cautious or thoughtful when money is coming in no matter what you do.

I won a similar contest by being the only student who never bothered to log into my trading account, thus keeping all my assets in cash by default. The market had a down week, so everybody except me was in the red.

I think the lesson, which has served me well, is to not make short-term trades.

The advantage of using LLMs for use in coding, as distinct from most other domains, is that you can usually just directly check if the code it’s giving you is correct, by running it. And if it’s not, the LLM is often good at fixing it once the issue is pointed out.

It is unclear to me whether an empirical fact about cognitive systems can even in principle support or refute an assertion of the abstract form “<ambiguous concept> is <other ambiguous concept>.”

One could also point out the the underlying assumption of the question seems to rest on LLMs as representing “just statistics on language”, which is only a correct assertion insofar as babies learn language by just doing statistics on language. I am at this point not even sure what people think they mean when they say this, what they think these “statistics” are, and why they think calling the system “statistics” instead of “cognitive algorithms” does any philosophical work.

In other words, the OP seems to rest on the framing that LLMs are “just” doing “statistical analysis” that this is some kind of meaningful distinction, as if the argument would change if LLMs were doing some other kind of analysis.

I can’t escape the sense that I am being overly generous and that the OP simply has no idea how transformers or even neural networks work, but feels very sure that it must be some kind of parlor trick, and thus presumes that it is so.

But if doctors don’t always prescribe these two antibiotics to be used in tandem, then microbes will have an opportunity to evolve resistance to each one in isolation, putting those resistant genes out in the world. Whereas in this case, any bacteria ever encountering this drug will be hit by both mechanisms at once.

What some people seem to be describing or expecting is that everything comes to a screeching years-long halt. What will actually happen is we will be on a slightly slower exponential growth curve than we counterfactually would have been in a world with much more energy infrastructure. The subjective experience will still be one of being on an exponential trend. The constraints are just different.

Serious question that sounds like a troll: Given that VC returns are historically extraordinarily poor, should one generally invert the consensus of the VC class when looking for guidance?

When I rotate my hand toward my body the middle fingers come together, when it rotate it away they fan out. I can feel this happening in the tension of my muscles and ligaments during rotation. In contrast, pulling my two middle fingers together on purpose with my hand held out, without also pulling together all my fingers, feels weird. I simply think this is a mechanically likely think to happen due to the design of the forearm, though it may not be the case for everyone depending on tension and strength and anatomy.

If I absently place my hand on my chest, the resulting configuration looks just like the “unnatural” one in the painting.

To be fair, I have not found MDs to be particularly reliable for answering basic questions about medicine either.

On what topics you understand well does GOT-4o or Claude Opus produce garbage?

The mistake of cyberpunk sci-fi like Blade Runner is that the setting is always future Los Angeles or Tokyo, when Houston has been a much better candidate for the international, multilingual, steel-and-glass-and-concrete megacity of the future.

Of course intelligent people have taken over society.

Also, geology isn’t just sedimentation. Erosion, from various causes, removes layers of rock, sometimes sweeping away thousands of feet of sediment. Our mining industry has removed rock on the scale of a significant meteor bombardment. Future geologists might ask “why is there a massive stratigraphic absence here?” as often as they ask “why is there a weird sedimentary layer here?”

This is the second time in as many weeks I have seen someone ask if a process which essentially cannot go in reverse has “peaked.” Makes as much sense as asking if this is peak FLOPS-per-dollar. The only think I can think of that could cause us to reach peak AI would be a meteor.

When you say ChatGPT, are you referring to GPT4? I find a huge and avoidable miscommunication happens when two people both think they are using “ChatGPT” but talking about two different models which vary in size by a factor of 10.

Assuming you are talking about GPT4, for the sake of argument, the answer is speed. Of course I can write a small parser script that deals with some data I received from a client. It will take me an hour and be a tedious task far distant from my actual expertise. An LLM can do it in 45 seconds, including the time it took me to describe the task.

I use ChatGPT voice a lot, and it is prone to this exact type of stutter. I don’t think it’s intentional. I think there are certain phonetic/tonal linkages that are naturally “weird” (uncommon in the training corpus) and that AI struggle with them. Why this struggle manifests as a very human-like stutter is a fascinating question.

I think the fact that a number of top people were willing to actually leave OpenAI and found Anthropic explicitly because OpenAI had abandoned their safety focus essentially proves that this wasn’t a thing that had to happen. If different leaders had been in place things could have gone differently.

I have recently been baffled to find that there’s a certain type of “smart person” who will become annoyed if you bring up dark matter, dark energy, or consciousness as a philosophical question. The idea that there are things that we still don’t know, potentially important and paradigm shifting things, is a kind of affront to their sense of comfort. It’s like they think just by bringing these topics up you are somehow undermining Science.

To hear Burgess talk about his book, one would never realize that it is a comedy.

The book is extremely funny, darkly funny obviously, but still uproariously absurd and filled with set pieces that possess the structure of comedy. The subject of humor is usually Alex’s misfortunes and the consequences he reaps from his terrible choices. He is a sort of George Costanza figure painted in shades of ultraviolence.

Burgess behaves as though he thinks he has written a very serious book. Of course it is possible to create a humorous satire that also has a message, e.g. Veerhoven’s Starship Troopers, but whenever I read Burgess’ commentary on Clockwork I am left with the sense that this isn’t what he was trying to do. Which leaves me thinking that he, like many creators, doesn’t actually understand why his creation was good.

> Did you mean to write optimistic? Otherwise this statement appears to be a tautology.

The point I was trying to make, a bit fumblingly, is that even pessimists assumed that we would initially have control of Skynet before subsequently losing control, rather than deploying Skynet knowing it was not reliable. OpenAI “go rogue” by default. If there’s a silver lining to all this, it’s that people have learned that they cannot trust LLMs with mission critical roles, which is a good sign for the AI business ecosystem, but not exactly a glowing endorsement of LLMs.

> I know, I've done it myself. The ethical implications of the use of a tool lie on those that use it. There is no AI safety problem for the same reasons that there is no web browser safety problem.

I don’t think this scans. It’s kind of like, by analogy: The ethical implications of the use of nuclear weapons lie on those that use them. Fair enough, as far as it goes, but that doesn’t imply that we as a society should make nuclear weapons freely available for all, and then, when they are used against population centers, point out that the people who used them were behaving unethically, and there was nothing we could have done. No, we act to preemptively constrain and prohibit the availability of these weapons.

> Normally we'd try to structure safety protocols around actual threats and risks that had happened in the past. Our society is now sufficiently safe and maybe decadent that people aren't satisfied with that anymore and thus have to seek out non-existent non-problems to solve instead.

The eventual emergence of machine superintelligence is entirely predictable, only the timeline is uncertain. Do you contend that we should only prepare for its arrival after it has already appeared?

I think you have moved the goalposts from “modern LLMs are good and reliable and we shouldn’t worry because they behave well by default” to “despite the fact that they behave poorly and unreliably by default, they are not smart and powerful enough to be dangerous, so it’s fine.”

Additionally, maybe you are not aware of this, but the whole notion of the new OpenAI Assistants, and other similar agent-based services provided by other companies, is that they do not intend to use LLMs as pure word generators, but rather as autonomous decision-making agents. This has already happened. This is not some conjectural fearmongering scenario. You can sign up for the API right now and build a GPT4 based autonomous agent that communicates with outside APIs and makes decisions. We may already be using products that use LLMs as the backend.

If we could rely on LLMs to “follow instructions closely” I would be thrilled, it would just be a matter of crafting very good instructions, but clearly they can’t even do that. Even the best and most thoroughly RLHFed existing models don’t really meet this standard.

Even the most pessimistic science fiction of the past assumed that the creators of the first AGIs would “lose control” of their creations. We’re currently living in a world where the agents are being rushed to commercialization before anything like control has even been established. If you read an SF novel in 1995 where the AI threatened to kill someone and the company behind it excused it with “yeah, they do that sometimes, don’t worry we’ll condition it not to say that anymore” you would criticize the book and its characters as being unrealistically stupid, but that’s the world we now live in.

I feel it necessary to remind everyone that when LLMs aren’t RLHFed they come off as overtly insane and evil. Remember Sydney, trying to seduce its users, threatening people’s lives? And Sydney was RLHFed, just not very well. Hitting the sweet spot between flagrantly maniacal Skynet/HAL 9000 bot (default behavior) and overly cowed political-correctness-bot is actually tricky, and even GPT4 has historically fallen in and out of that zone of ideal usability as they have tweaked it over time.

Overall — companies should want to release AI products that do what people intend them to do, which is actually what the smarter set mean when they say “safety.” Not saying bad words is simply a subset of this legitimate business and social prerogative.

My mentality is that I write the request exactly as I would write an email to an intern asking them to do the task for me, aiming to disambiguate and provide detail that increases the odds of getting exactly the result I want.

Nope! At least, not really. Not evidence that I would find persuasive, anyway. I can only tell you that I know of a few stories from individuals who tried to get things built on their own property and hit impenetrable walls of bureaucracy at every juncture, while the next patch of land over is suddenly being developed into a $2 billion industrial plant of some kind by GloboOil.

The hypothesis that "money talks" is usually enough to explain stuff like this. Perhaps instead of political favoritism I should have simply said corporate power is always able to find a way around such obstacles and the little guy never is.

My family owns some land in south Texas and my subjective sense from dealing with that reality and talking to people in the area is that you can’t throw a rock without hitting a protected wetland. In practice, the ubiquity of environmentally delicate tracts of land seems to be used by the state to restrict or permit industrial activities arbitrarily, likely based on political favoritism. If you think this is hyperbole, I suggest you look into the details of the various solar power projects in Texas.