Recent investigations into AI reveal that comedians and writers who excel at clever wordplay might find temporary solace.
Researchers from institutions in the UK and Italy have been exploring the capacity of large-scale language models (LLMs) to comprehend puns, only to discover significant gaps in their understanding.
For instance, they examined the statement, “I used to be a comedian, but my life became a joke.” Even after substituting it with “I used to be a comedian and my life became a mess,” LLMs still acknowledged the presence of puns.
Another example tested was: “long fairy tales have a tendency to dragonify.” When “dragon” was swapped with its synonym “extension” or any arbitrary word, LLMs erroneously assumed a pun was present.
Professor Jose Camacho Collados, associated with Cardiff University’s School of Computer Science and Informatics, suggested that the research indicates a fragile understanding of humor by LLMs.
“Essentially, LLMs tend to retain information from their training, allowing them to recognize established puns, but that doesn’t equate to true understanding,” he remarked.
“We consistently managed to mislead the LLM by altering existing puns and stripping away the double meanings integral to the original humor. In these scenarios, the model would draw connections to prior puns and create various justifications for its conclusions. Ultimately, we determined that the model’s interpretation of puns was merely an illusion.”
The findings indicated that LLMs’ accuracy in differentiating between pun and non-pun sentences could dip to 20% when encountering unfamiliar wordplay.
Another pun tested was: “Old LLM never dies, it just loses attention.” Even when attention shifted to “ukulele,” the LLM still identified it as a pun, as “ukulele” bore a slight resemblance to “you-kill-LLM.”
The team was impressed by the creativity displayed, yet the LLM still failed to appreciate the humor.
The researchers emphasized that their findings underscore the need for caution when utilizing LLMs for tasks that involve humor, empathy, and an understanding of cultural subtleties.
Over a century has passed since the Endurance, reputedly the strongest wooden ship ever constructed, met its fate in Antarctic ice. Recent evaluations of historical evidence indicate that it might not have been as resilient as other polar vessels of its era, casting doubt on expedition leader Ernest Shackleton’s awareness of its limitations.
Shackleton aimed to journey across Antarctica from the Weddell Sea to the Ross Sea, making stops along the way. However, the Endurance never reached the Antarctic coastline. In 1915, it became trapped in ice in the Weddell Sea, leading to its sinking.
Jukka Tukuri, who was part of a significant expedition at Aalto University in Finland, discovered a shipwreck on the seabed in 2022. Upon researching the polar vessels of that time, he found that the narrative surrounding the Endurance was misleading; it lacked the structural strength attributed to it.
During the late 19th and early 20th centuries, several ships were engineered specifically to navigate sea ice. Many featured an oval hull shape and shallow keels, which helped prevent ice from creating strong hold on the side, allowing it to slide beneath. Additionally, the internal structure of these vessels included a continuous deck spanning the hull’s entire length, which enhanced their rigidity.
In contrast, the Endurance was longer and had taller keels. Tukuri’s calculations indicated that other contemporary polar ships could withstand 1.7 to 2.7 times more compressive force than the Endurance. Furthermore, the ship’s large engine restricted the lower deck to only part of the vessel, creating structural weaknesses due to the absence of a reinforced box-like configuration.
In his review of Shackleton’s letters, Tukuri found evidence that Shackleton was aware of these design flaws. Just before departing for Antarctica, Shackleton noted in a letter to his wife that Nimrod, his previous ship, was more robust. Yet, he pressed on with his journey: “He was willing to embrace the risk,” Tukuri commented.
As predicted, the Endurance could not withstand the immense pressure of the sea ice. The vessel was compressed and distorted until its keel was ultimately torn apart and submerged.
Despite this, a myth arose portraying the Endurance as the world’s most resilient wooden ship, potentially fueled by an article in The Era. According to Tukuri, Shackleton perpetuated this narrative; the reasons remain unclear, but it seems to add an element of drama to his ill-fated expedition. “Endurance may have symbolized strength and heroism poetically,” he said. “Regrettably, that was not the case from an engineering perspective.”
What actions can the UK government take regarding Twitter? Should What are your thoughts on Twitter? What interests does Elon Musk have?
The billionaire proprietor of the social network, still officially referred to as X, has had an eventful week causing disruptions on his platform. Besides his own posts, which include low-quality memes sourced from 8chan and reposted fake concerns from far-right figures, the platform as a whole, along with the other two of the three “T’s,” TikTok and Telegram, briefly played a significant role in orchestrating this chaos.
There is a consensus that action needs to be taken: Bruce Daisley, former VP EMEA at Twitter, proposes individual accountability.
In the near term, Musk and other executives should be reminded of their legal liability for their actions under current laws. The UK’s Online Safety Act 2023 should be promptly bolstered. Prime Minister Keir Starmer and his team should carefully consider if Ofcom, the media regulator frequently criticized for the conduct of organizations like GB News, can effectively manage the rapid behavior of someone like Musk. In my view, the threat of personal consequences is much more impactful on corporate executives than the prospect of a corporate fine. If Musk continues to incite unrest, an arrest warrant could create sparks from his fingertips, though as a jet-setting personality, an arrest warrant could be a compelling deterrent.
Last week, London Mayor Sadiq Khan presented his own suggestion.
“The government swiftly realized the need to reform the online safety law,” Khan told the Guardian in an interview. “I believe that the government must ensure that this law is suitable immediately. I don’t think it currently is.”
“Responsible social media platforms can take action,” Khan remarked, but added that “if they fail to address their own issues, regulation will be enforced.”
When I spoke to Euan McGaughey, a law professor at King’s College London on Monday, he provided more precise recommendations on what actions the government could take. He mentioned that the Communications Act 2003 underlies many of Ofcom’s authorities and is applied to regulate broadcast television and radio, but extends beyond those media.
Simply as section 232 specifies that “television licensable content services” involve distribution “by any means involving the use of an electronic communications network,” this Act empowers Ofcom to regulate online media content. While Ofcom could exercise this power, it is highly improbable as Ofcom anticipates challenges from tech companies, including those fueling riots and conspiracy theories.
Even if the BBC or the government were reluctant to interpret the old law differently, minor modifications could subject Twitter to stricter broadcasting regulatory oversight, he added.
For instance, there is no distinction between Elon Musk posting a video on X about (so-called) two-tier policing, discussing “detention camps” or asserting “civil war is inevitable” and ITV, Sky, or the BBC broadcasting the news… Online Safety Act Grossly insufficient, as the constraints merely aim to prevent “illegal” content and do not inherently address false or dangerous speech.
The law of keeping promises
Police in Middlesbrough responded to a mob spurred by social media posts this month. Photo: Gary Culton/Observer
It may seem peculiar to feel sympathy for an inanimate object, but the Online Safety Act has likely been treated quite harshly given its minimal enforcement. A comprehensive law encompassing over 200 individual clauses, it was enacted in 2023, but most of its modifications will only take effect once Ofcom has completed the extensive consultation process and established a code of practice.
The law introduces a few new offenses, such as bans on cyber-flashing and upskirt photography. Sections of the old law, referred to as malicious communications, have been substituted with new, more precise laws like threatening and false communications, with two of the new offenses going into effect for the first time this week.
But what if this had all happened earlier and Ofcom was operational? Would the outcome have been different?
The Online Safety Act is a peculiar piece of legislation: an effort to curb the worst impulses on the internet, drafted by a government taking a stance in favor of free speech amidst a growing culture war and enforced by regulators staunchly unwilling to pass judgment on individual social media posts.
What transpired was either a skillful act of navigating a tricky situation or a clumsy mishap, depending on who you ask. The Online Safety Act does not outright criminalize everything on the web; instead, it mandates social media companies to establish specific codes of conduct and consistently enforce them. For certain forms of harm like incitement to self-harm, racism, and racial hatred, major services must at least provide adults with the option to opt out of such content and completely block it from children. For illegal content ranging from child abuse imagery to threats and false communications, it requires new risk assessments to aid companies in proactively addressing these issues.
It’s understandable why this legislation faced significant backlash upon its passage: its main consequence was a mountain of new paperwork in which social networks had to demonstrate adherence to what they had always purportedly done: attempting to mitigate racist abuse, addressing child abuse imagery, enforcing their terms of use, and so forth.
Advocates of the law argue that it serves more as a means for Ofcom to impose its promises on companies rather than forcing them to alter their behavior. The easiest way to impose a penalty under the Online Safety Act – potentially amounting to 10% of global turnover if modeled after GDPR – is to announce loudly to customers that steps are being taken to tackle issues on the platform, only to do nothing.
One could envision a scenario where the CEO of a tech company, the key antagonist in this play, stands before an inquiry, solemnly asserting that the reprehensible behavior they witness violates their terms of service, then returning to their office and taking no action.
The challenge for Ofcom lies in the fact that multinational social networks are not governed by cartoonish villains who flout legal departments, defy moderators, and whimsically enforce one set of terms of service on allies and a different one on adversaries.
Except for one.
Do as I say, don’t do as I do
Elon Musk’s Twitter has emerged as a prime test case for online safety laws. On the surface, the social network appears relatively ordinary: its terms of service prohibit the dissemination of much of the same content as other major networks, with a slightly more lenient stance on pornographic material. Twitter maintains a moderation team that employs both automated and human moderation to remove objectionable content, an appeals process for individuals alleging unfair treatment, and progressive penalties that could ultimately lead to account suspensions for violations.
However, there’s an additional layer to how Twitter operates: Elon Musk follows through on what he says. For instance, last summer, after a prominent right-wing influencer shared child abuse images, the account’s creator received a 129-year prison sentence. The motive remains unclear, but the account was swiftly suspended. Musk then intervened:
The only people who have seen these photos are members of the CSE team. At this time, we will remove these posts and reinstate your account.
While Twitter’s terms of service theoretically prohibit many of the egregious posts related to the UK riots, such as “hateful conduct” and “inciting, glorifying, or expressing a desire for violence,” they do not seem to be consistently enforced. This is where Ofcom may potentially take aggressive actions against Musk and his affiliated companies.
If you wish to read the entire newsletter, subscribe to receive TechScape in your inbox every Tuesday.
It’s a common belief that it requires 10,000 hours of practice to master a skill, but studies suggest that not everyone possesses the innate talent required to become an Olympian or Paralympian. While practice can enhance performance, genetic factors impacting both physical strength and cognitive abilities likely distinguish between “good” and “great” athletes.
The success in sports is also influenced by factors beyond one’s control, such as birth timing. For instance, in the 2010-11 UEFA Youth Football Tournament, 43% of players were born between January and March (earlier in the selection period). Only 9% of players were born between October and December.
Older children who start school may have a physical advantage over their younger peers in terms of size, strength, and confidence. However, many sports psychologists argue that any birth month advantage is also influenced by social factors, such as how teachers perceive a child’s abilities.
Therefore, the most effective approach to becoming an Olympian may involve exploring a range of interests as a child and then focusing on activities where natural talent and enjoyment are evident.
Meta has lifted previous restrictions on Donald Trump’s Facebook and Instagram accounts as the 2024 presidential election approaches, the company announced on Friday.
After being banned for his online behavior during the January 6 riot, President Trump was allowed to return to the social network in 2023 with “guardrails” in place. But those guardrails have now been removed.
“In assessing the responsibility of permitting political expression, I believe the American people should be able to hear from presidential candidates with the same standards,” Mehta said in a blog post, alluding to Trump formally becoming the party’s nominee at the Republican National Convention scheduled for next week.
As a result, Mr. Trump’s account will no longer be subject to the harsh suspension, which he said was instituted in response to “extreme and extraordinary circumstances” and “was not necessary to apply.”
“All US presidential candidates are required to follow the same community standards as all Facebook and Instagram users, including policies to prevent hate speech and incitement to violence,” the company said in a blog post.
Since returning to the meta social network, Trump has mainly used his account to share campaign information, attacks on Democratic candidate Biden and memes.
Critics of Trump and online safety advocates have expressed concern that his return could lead to an increase in misinformation and incitement to violence like that seen during the storming of the Capitol, which initially prompted the president’s travel ban.
The Biden campaign condemned Mehta’s decision in a statement on Friday, calling it a “greedy and reckless decision” that amounts to “a direct attack on our security and democracy.”
“Restoring his access would be like giving car keys to someone you know is going to drive his car into a crowd and off a cliff,” campaign spokesman Charles Kretschmer Luttwak said. “It’s like giving a megaphone to a real racist who is going to shout hatred and white supremacy from the rooftops and make it mainstream.”
In addition to the Meta platform, other major social media companies, including Twitter (now X), Snapchat and YouTube, have also banned Trump’s accounts due to his online activity surrounding the January 6 attack.
The former president was allowed to return to X last year following a decision by Elon Musk, who bought the company in 2022, but has yet to tweet.
Trump Came back It is set to appear on YouTube in March 2023. He remains banned from Snapchat.
Trump launched his own social network, Truth Social, in early 2022.
It’s important to have friends who come to your birthday parties, offer support during tough times, and allocate resources to develop apps for emerging virtual reality platforms despite limited direct benefits. It may be tempting to believe that a $30 billion cash reserve and a product line generating over $200 billion annually are sufficient. However, Apple is finding that money cannot buy everything.
Pre-orders for Apple’s Vision Pro headset, a $3,500 “spatial computing” platform and CEO Tim Cook’s vision of Apple’s future, opened last week. Despite Apple’s enthusiasm, quiet opposition from potential users has overshadowed the announcement.
According to a report from Bloomberg (£), Netflix has opted not to design a Vision Pro app or support existing iPad apps on the platform, instead instructing users to access their content through a web browser.
Rather than developing a Vision Pro app or supporting existing iPad apps, Netflix has chosen to direct users to watch their content on the web. This decision is notable given the competition between Netflix and Apple in the streaming market.
Although the initial weekend release of Vision Pro saw an estimated 160,000-180,000 units sold, this pales in comparison to Netflix’s 250 million paying subscribers. Therefore, Netflix’s reluctance to invest resources in an app for the Vision Pro is understandable, as app development is only worthwhile if it can attract new customers or retain existing ones.
Despite Apple’s promotion of the Vision Pro as the most immersive way to watch TV, Netflix has similarly abandoned its app for MetaQuest, demonstrating a pattern of resistance to immersive platforms.
Due to these decisions, Vision Pro users will be limited to watching Netflix through the web, losing the ability to access offline viewing, a key selling point of the headset.
Furthermore, YouTube and Spotify have also opted not to release new apps for the Vision Pro, indicating a lack of enthusiasm from major content providers for the platform.
In a related story, Apple has recently allowed developers to bypass its payment system, providing them with an alternative to the high fees associated with in-app purchases. This shift may reflect a broader resistance among developers to Apple’s monopoly over economic activity in their app ecosystem.
The reluctance of major content providers to invest in apps for the Vision Pro may indicate a broader skepticism among developers about the benefits of supporting Apple’s latest venture. This trend may signal a greater movement within the developer community to challenge Apple’s control over app development and monetization.
This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.
Strictly Necessary Cookies
Strictly Necessary Cookie should be enabled at all times so that we can save your preferences for cookie settings.
