Sort:  

Let us now summarize how the model of government described in the Constitution compares to the Roman Republic:

Two houses in the legislature – both.
A senior body of experienced men and a junior body close to the people – both
Senators elected by the people - neither
A senior magistrate as executive – United States one; Rome two.
One male landholders can vote – both

At the time of the Constitution’s ratification, our government was as close to the Roman Republic as it would ever be. And then things began to change as we moved in the direction of a democracy. The property qualification was steadily reduced until about 1850 when it was removed completely; the electors came to be chosen by the people instead of the state legislatures, giving the people a direct say in electing the president; and finally, in 1910, the law for electing senators was changed to allow the people to elect them directly.

At the time of the Constitution’s ratification, our government was as close to the Roman Republic as it would ever be. And then things began to change as we moved in the direction of a democracy. The property qualification was steadily reduced until about 1850 when it was removed completely; the electors came to be chosen by the people instead of the state legislatures, giving the people a direct say in electing the president; and finally, in 1910, the law for electing senators was changed to allow the people to elect them directly.

The differences between the two Republics was dictated by the difference in culture and time, but our attraction to the ancient system because it had no monarch led us to the creation of a new version designed to withstand the modern age.

Ivanka Trump was born on October 30, 1981, in Manhattan, New York, to real estate mogul Donald Trump and former model Ivana Trump. Growing up in a high-profile family, she was accustomed to the public eye from an early age. After her parents' divorce when she was 10, Ivanka attended several prestigious institutions. She first studied at the Chapin School in New York City and later transferred to Choate Rosemary Hall, a renowned boarding school in Connecticut. Although unhappy in her new environment, she made a commitment to maintain strong academic performance.

Ivanka Trump is well-known for her diverse pursuits beyond her family's real estate empire. In 2009, she released a New York Times best-selling book titled "The Trump Card: Playing to Win in Work and Life", establishing herself as a voice for the professional millennial woman. Ivanka also launched her fashion and lifestyle brand, the Ivanka Trump Collection, which reflected her dedication to empowering women through work and style. However, she stepped back from the business side in 2017 to focus on her role as a senior adviser in her father’s administration, ultimately announcing the closure of her brand to concentrate fully on her political work.

How did Ivanka Trump contribute to politics?

Ivanka Trump played a crucial role in her father's 2016 presidential campaign, publicly supporting him and introducing him at the Republican National Convention. As a senior advisor during her father's presidency, she focused on issues such as women's empowerment and family policies, and was involved in initiatives like the Paycheck Protection Program during the COVID-19 pandemic.

What is Ivanka Trump's family life like?

Ivanka Trump married real estate developer Jared Kushner in 2009, and they have three children: Arabella, Joseph, and Theodore. She converted to Orthodox Judaism and observes the faith's traditions, including adhering to a kosher diet. Ivanka's family life is enriched by her dedication to both her role as a mother and her involvement in public service.

I studied CS and Psych. You are correct about how we dont understand the brain, but even what we do understand isnt replicable in a data center.
There are two points to this- 1. the brain operates not just on electric style neural networks, but also on waves, chemicals (solids, liquids, gases, probably plasma), and "quantum" phenomena we hardly understand. it'll never get beyond an abstraction, never faithfully replicate human cognition. 2. there is no intelligent brain without a body - to understand the world you have to move, grow, eat, touch, communicate, etc. AGI in a robot body would be more likely than a datacenter, but even then it doesnt really grow, interact, learn, etc the same ways a human/animal does, and therefore its knowledge will be different.

So, you're missing one of the most important parts of why Deepseek is powerful. It's the reinforcement loop. They discovered and open sourced the o1, o3 test time compute methodology (or at least something comparable), that uses reinforcement learning to train an LLM to generate a chain of thought process that maximizes the reward of getting the right answer (or process if it's a process rather than an outcome reward). This is the thing that can potentially allow an LLM to get above human level performance in the same way that Alphazero could exceed human Go players. One thing that people don't appreciate is that you could theoretically train an LLM with reinforcement learning alone, and no human training data.

It's just that it would take an obscenely long time to get any signal out of the random generations of the LLM, which is why we start it out with pretraining and fine tuning. But it's good to recognize this, and not get stuck in this thinking that these models can only get as good as the human training data that went in. This is only true of the pretraining and fine tuning steps.

To give an example, we could use this reinforcement loop to train an LLM not only to give the right answer to a coding problem, but do it in the most concise way, or with the shortest execution time, or smallest memory footprint, etc. Eventually, the LLM would drift away from the "human" way of working through code and invent it's own process that humans might not even find recognizable. I'm not suggesting we do this, it's just an example of where things are going, and why Deepseek is much more interesting that just MoE and distillation.

Hi, @taskmaster4450le,

This post has been voted on by @darkcloaks because you are an active member of the Darkcloaks gaming community.


Get started with Darkcloaks today, and follow us on Inleo for the latest updates.

Zuckerberg is willing to take on such risks because of the potential benefits that come with smoothing his relationship with Trump, the people familiar with the matter said. In contrast to Meta's previous strategy of advocating for an even playing field across the tech industry, Zuckerberg now sees opportunities to gain a strategic advantage for his company, the people said.

One major concern for Zuckerberg is Elon Musk's central position in the Trump administration, where he's focused on slashing regulations. Meta competes with Musk's X and is also investing heavily in artificial intelligence, an area of particular interest to Musk and his startup xAI. Musk's role in the White House could put Meta at a disadvantage when it comes to policies surrounding AI.

Using a thrusting sword in close quarters was described on a least two other occasions: against the Gauls in 387 B.C. as described by Plutarch and from Polybius describing Cannae and Telamon. In these cases, the author tells how the Gallic blades were inferior because of the quality of the metal and the lack of a point on the sword. The Romans had learned how to avoid the swings of the Gauls and stab them below the shield. The Gauls could only strike one or two blows before their swords bent and they would have to stomp on them to straighten them out.

A confusing story about the origin of the Gladius is given to us by Polybius. Quoting from a fragment of book twenty-two, “The Celtiberians excel the rest of the world in the construction of their swords; for their point is strong and serviceable, they can deliver a cut with both edges. Wherefore the Romans abandoned their traditional swords after the Hannibalian War and adopted those of the Iberians. They adopted, I say, the construction of the swords, but they can, by no means imitate the excellence of the steel or the other points in which they are elaborately finished.”

What was this ancestral sword they abandoned? Can we assume a pointed gladius did not exist before that time? Contradicting himself, Polybius told the story of an engagement by Flaminius in 223 B.C. where the Gallic swords are mentioned as having no points. This implies the Roman swords did have points.

It’s interesting to look at the metallurgy of Gladii found in Europe. They are mostly wrought iron with carbon content at .03%. The edges were sharpened by forging (hammering) or sharpening on a wheel. Most were fabricated by placing strips of iron together in a sandwich. The quality is variable probably due to the skill of European smiths of the time.

The Romans, for the most part, used local smiths to fabricate swords for the army. Tacitus cites a case where Vespasian commanded that “strong cities be picked out so as to get their arms factories busy.” One of the more famous sword factories was located at Rheims.

The second area of discussion is casualties. Data shows that casualties among armies Rome defeated amount to about 50%. This contrasts with the figure for hoplite battles which ran about 14%. The difference is explained by pursuit which occurred more often in Roman times. One has a hard time imaging a hoplite phalanx dragging their heavy shields in pursuit of a retreating enemy. The Romans, however, made a practice of it. When Rome was a victor, it lost about 5% of its force, although there was the occasional “Pyrrhic victory”. So the typical Roman battle had low or high casualties on each side until the rout when the losing side’s casualties accelerated.

Another factor in battle is mobility. Opposing forces would advance or give ground depending on the circumstances. We have the famous story of the ordered Carthaginian pull back at Cannae creating a “crescent” for the Romans to advance into. Caesar cited a case where the Helvetii fell back a mile before resuming their resistance from a higher position. In sum, mobility had the impact of influencing battle conditions depending on the terrain or composition of the armies. The impact of cavalry is well documented.

Kaplan's ascendency at Meta coupled with the company's policy changes has solidified a political shift to the right, multiple sources said.

Since joining Meta in 2011 as a policy vice president, Kaplan has built a reputation as an executive who takes calculated risks even if it means upsetting some people internally, the people said.

In 2018, Kaplan made headlines for attending Brett Kavanaugh's highly contentious Supreme Court confirmation hearing as a personal friend. His appearance caused so much controversy that Meta was forced to address the matter in a statement, saying the "leadership team recognizes that they've made mistakes handling the events of the last week and we're grateful for all the feedback from our employees."

Meta in January agreed to pay $25 million as part of a settlement with Trump over the company's decision to suspend his accounts following the Capitol riot. In January 2023, Meta said it was reinstating Trump on its platform after the two-year suspension.

The company's efforts to win favor with Trump seem to be working, at least based on what the president has publicly said.

After Kaplan announced Meta's major content-moderation and related policy shifts in early January on "Fox and Friends," Trump appeared to be impressed.

The app came back online after President Donald Trump said he would postpone enforcement of the ban. He signed an executive order on his first day in office to extend the law's deadline by an additional 75 days to April 5.

Trump said on his Truth Social platform that he "would like the United States to have a 50% ownership position in a joint venture" in order to "save TikTok, keep it in good hands and allow it to say up."

Despite being removed from the two app stores for nearly a month, TikTok had recovered about 90% of the traffic it was seeing prior to the law's Jan. 19 deadline, CNBC reported in late January, citing data from Cloudflare Radar.

Airbnb shares pop 15% on better-than-expected earnings and revenue
Airbnb reported fourth-quarter earnings on Thursday that beat analysts' estimates, but the company issued soft guidance.

Airbnb shares jumped 15% in extended trading on Thursday after the company reported fourth-quarter earnings and revenue that beat analysts' estimates.

Here's how the company did compared with what Wall Street was expecting, based on a survey of analysts by LSEG:

Earnings per share: 73 cents vs. 58 cents expected
Revenue: $2.48 billion vs. $2.42 billion expected
Revenue increased 12% from $2.22 billion during the same period last year. Airbnb reported net income of $461 million, or 73 cents per share. A year ago, Airbnb reported a loss of $349 million, or a loss of 55 cents per share.

In a letter to shareholders, the company said it has rebuilt its technology stack and rolled out new product updates in recent years that are driving higher conversion rates. The company also teased new offerings that it will launch in May, though it did not disclose additional details.

"Airbnb is a fundamentally stronger company today than it was several years ago," the letter said. "We're continuing to build on this momentum in 2025, executing a multi-year strategy to perfect the core service, accelerate growth in global markets, and launch and scale new offerings."

The company reported 111 million nights and experiences booked, up 12% from a year ago and above the 108.7 million expected by StreetAccount. Gross booking value, which Airbnb uses to report host earnings, service fees, cleaning fees and taxes, totaled $17.6 billion in the fourth quarter. That is above the $17.2 billion expected by analysts polled by StreetAccount.

Airbnb said Airbnb.org, a nonprofit founded by the company, has housed more than 19,000 people and 2,300 pets following the devastating wildfires that broke out in Los Angeles last month. The nonprofit has pledged to offer 100,000 free nights and has received $27 million in donations, including $18 million from Airbnb and its founders.

The Roman army used the phalanx as a attack formation until they realized the futility of the phalanx on un-level ground during a battle with the Samnites in 315 B.C. The consuls decided to create a more flexible system as a result the manipular formation was born. Later, the manipular system was replaced by the cohort system.

The drawing above compares the manipular system with the cohort system.

In Republican times, the Roman army marched toward the enemy in three lines Hastati, Principes, and Triarii. A legion would contain 10 maniples of each of the three types as its infantry component.

The cohort system was adopted at least partly in response to the difficulty fighting barbarians. The Celts and Spaniards were able to defeat a highly trained army because of their concentrated charges. The way to counter them was to concentrate men, in formation, on the Roman side: building cohorts out of maniples.

The company reported revenue of $469 million in 2021, representing 213% year-over-year growth due to a combination of factors, including a boost from the COVID-19 pandemic, according to regulatory filings. Revenue popped again in 2022 to $746.6 million, but at a lower 59% year-over-year growth compared to the prior year. Revenue increased in 2023 to $879.7 million, representing 18% year-over-year growth.

The company has been profitable since 2022 on net income at $154.7 million, although that result dipped to $14.7 million in 2023. The company has not reported full-year results for 2024.