Valve is a unique company with no traditional hierarchy. In business school, I read a very interesting Harvard Business Review article on the subject. Unfortunately it’s locked behind a paywall, but this is Google AI’s summary of the article which I confirm to be true from what I remember:

According to a Harvard Business Review article from 2013, Valve, the gaming company that created Half Life and Portal, has a unique organizational structure that includes a flat management system called “Flatland”. This structure eliminates traditional hierarchies and bosses, allowing employees to choose their own projects and have autonomy. Other features of Valve’s structure include:

• Self-allocated time: Employees have complete control over how they allocate their time
• No managers: There is no managerial oversight
• Fluid structure: Desks have wheels so employees can easily move between teams, or “cabals”
• Peer-based performance reviews: Employees evaluate each other’s performance and stack rank them
• Hiring: Valve has a unique hiring process that supports recruiting people with a variety of skills

Someone did the math and realized we would need a 130% tariff on all goods to replace current income tax revenue.

People’s number one concern is inflation. If that tariff is created we will see 100% inflation over night!

You do realize that every posted on the Fediverse is open and publicly available? It’s not locked behind some API or controlled by any one company or entity.

Fediverse is the Wikipedia of encyclopedias and any researcher or engineer, including myself, can and will use Lemmy data to create AI datasets with absolutely no restrictions.

To add to this insight, there are many recent publications showing the dramatic improvements of adding another modality like vision to language models.

While this is my conjecture that is loosely supported by existing research, I personally believe that multimodality is the secret to understanding human intelligence.

I am an LLM researcher at MIT, and hopefully this will help.

As others have answered, LLMs have only learned the ability to autocomplete given some input, known as the prompt. Functionally, the model is strictly predicting the probability of the next word⁺, called tokens, with some randomness injected so the output isn’t exactly the same for any given prompt.

The probability of the next word comes from what was in the model’s training data, in combination with a very complex mathematical method to compute the impact of all previous words with every other previous word and with the new predicted word, called self-attention, but you can think of this like a computed relatedness factor.

This relatedness factor is very computationally expensive and grows exponentially, so models are limited by how many previous words can be used to compute relatedness. This limitation is called the Context Window. The recent breakthroughs in LLMs come from the use of very large context windows to learn the relationships of as many words as possible.

This process of predicting the next word is repeated iteratively until a special stop token is generated, which tells the model go stop generating more words. So literally, the models builds entire responses one word at a time from left to right.

Because all future words are predicated on the previously stated words in either the prompt or subsequent generated words, it becomes impossible to apply even the most basic logical concepts, unless all the components required are present in the prompt or have somehow serendipitously been stated by the model in its generated response.

This is also why LLMs tend to work better when you ask them to work out all the steps of a problem instead of jumping to a conclusion, and why the best models tend to rely on extremely verbose answers to give you the simple piece of information you were looking for.

From this fundamental understanding, hopefully you can now reason the LLM limitations in factual understanding as well. For instance, if a given fact was never mentioned in the training data, or an answer simply doesn’t exist, the model will make it up, inferring the next most likely word to create a plausible sounding statement. Essentially, the model has been faking language understanding so much, that even when the model has no factual basis for an answer, it can easily trick a unwitting human into believing the answer to be correct.

—-

⁺more specifically these words are tokens which usually contain some smaller part of a word. For instance, understand and able would be represented as two tokens that when put together would become the word understandable.

Agreed.

Nevertheless, the Federal regulators will have an uphill battle as mentioned in the article.

Neither “puffery” nor “corporate optimism” counts as fraud, according to US courts, and the DOJ would need to prove that Tesla knew its claims were untrue.

The big thing they could get Tesla on is the safety record for autosteer. But again there would need to be proof it was known.

I am a pilot and this is NOT how autopilot works.

There is some autoland capabilities in the larger commercial airliners, but autopilot can be as simple as a wing-leveler.

The waypoints must be programmed by the pilot in the GPS. Altitude is entirely controlled by the pilot, not the plane, except when on a programming instrument approach, and only when it captures the glideslope (so you need to be in the correct general area in 3d space for it to work).

An autopilot is actually a major hazard to the untrained pilot and has killed many, many untrained pilots as a result.

Whereas when I get in my Tesla, I use voice commands to say where I want to go and now-a-days, I don’t have to make interventions. Even when it was first released 6 years ago, it still did more than most aircraft autopilots.

I’m an AI researcher at one of the world’s top universities on the topic. While you are correct that no AI has demonstrated self-agency, it doesn’t mean that it won’t imitate such actions.

These days, when people think AI, they mostly are referring to Language Models as these are what most people will interact with. A language model is trained on a corpus of documents. In the event of Large Language Models like ChatGPT, they are trained on just about any written document in existence. This includes Hollywood scripts and short stories concerning sentient AI.

If put in the right starting conditions by a user, any language model will start to behave as if it were sentient, imitating the training data from its corpus. This could have serious consequences if not protected against.

Not quite, this was made with a ControlNet. A hybrid image wouldn’t work as well as this does. But the underlying visual phenomena is the same.

This is done by combining a Diffusion model with ControlNet interface. As long as you have a decently modern Nvidia GPU and familiarity with Python and Pytorch it’s relatively simple to create your own model.

The ControlNet paper is here: https://arxiv.org/pdf/2302.05543.pdf

I implemented this paper back in March. It’s as simple as it is brilliant. By using methods originally intended to adapt large pre-trained language models to a specific application, the author’s created a new model architecture that can better control the output of a diffusion model.

I am a satellite software engineer turned program manager. This is not unexpected in this current environment, however the conditions that created the environment are abnormal.

This solar cycle is much stronger than past cycles. I’m on mobile, so I can’t get a good screenshot, but you can go here to see this cycle and the last cycle, as well as an overlay of a normal cycle https://www.swpc.noaa.gov/products/solar-cycle-progression

As solar flux increases, the atmosphere expands considerably, causing more drag than predicted. During periods of solar minimum, satellites can remain in a very low orbit with minimal station keeping. However, at normal levels of solar maximum, 5 year orbits can easily degrade to 1 year orbits. Forecasters says we are still a year away from solar maximum, and flux is already higher than last cycle’s all time high (which was also an anomalously strong cycle). So it will get worse before it gets better.

TLDR: Satellites are falling out of the sky because the sun is angy