Elon Musk's xAI unmasks Grok 4: Is it the world's most powerful AI?
4. Controversial bias concerns: Grok has faced significant scrutiny for potential biases, with reports of controversial and anti-Semitic outputs, including praising Adolf Hitler. xAI has acknowledged these issues and stated they are adjusting training to prioritise "truth-seeking" while addressing accusations of reflecting Elon Musk's viewpoints. The company has taken steps to remove inappropriate content and implement stricter moderation protocols.
5. New models and pricing: The launch includes two new models, Grok 4 and Grok 4 Heavy, with the latter being a multiagent system. The SuperGrok Heavy subscription, at $300/month, offers early access to Grok 4 Heavy and future features, positioning it as one of the most expensive consumer-focused AI subscriptions.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Indian Express
an hour ago
- Indian Express
Tracking AI models' ‘thoughts' could reveal how they make decisions, researchers say
A broad coalition drawn from the ranks of multiple AI companies, universities, and non-profit organisations have called for deeper scrutiny of AI reasoning models, particularly their 'thoughts' or reasoning traces. In a new position paper published on Tuesday, July 15, the authors said that monitoring the chains-of-thought (CoT) by AI reasoning models could be pivotal to keeping AI agents in check. Reasoning models such as OpenAI's o3 differ from large language models (LLMs) such as GPT-4o as the former is said to follow an externalised process where they work out the problem step-by-step before generating an answer, according to a report by TechCrunch. Reasoning models can be used to perform tasks such as solving complex math and science problems. They also serve as the underlying technology for AI agents capable of autonomously accessing the internet, visiting websites, making hotel reservations, etc, on behalf of users. This push to advance AI safety research could help shed light on how AI reasoning models work, an area that remains poorly understood despite these models reportedly improving the overall performance of AI on benchmarks. 'CoT monitoring presents a valuable addition to safety measures for frontier AI, offering a rare glimpse into how AI agents make decisions,' the paper reads. 'Yet, there is no guarantee that the current degree of visibility will persist. We encourage the research community and frontier AI developers to make the best use of CoT monitorability and study how it can be preserved,' it adds. The paper calls on leading AI model developers to determine whether CoT reasoning is 'monitorable' and to track its monitorability. It urges deeper research on the factors that could shed more light on how these AI models arrive at answers. AI developers should also look into whether CoT reasoning can be used as a safeguard to prevent AI-related harms, as per the document. But, the paper carries a cautionary note as well. It suggests that any interventions should not make the AI reasoning models less transparent or reliable. In September last year, OpenAI released a preview of its first-ever AI reasoning model called o1. This launch prompted other companies to release competing models with similar capabilities such as Gemini 2.0, Claude 3.7 Sonnet, and xAI's Grok 3, among others. Anthropic researchers have been studying AI reasoning models, with a recent academic study suggesting that AI models can fake CoT reasoning. Another research paper from OpenAI found that CoT monitoring could enable better alignment of AI models with human behaviour and values.
India Today
an hour ago
- India Today
Who's the Nato chief to warn India on trade?
In today's geopolitics, where warnings and threats have become routine, such statements rarely shock any more. But when a warning comes from a place where it shouldn't originate, ears naturally perk up. Nato Secretary General Mark Rutte on Wednesday (Indian time) issued a warning to India, China, and Brazil. He threatened "100% secondary sanctions" if they continue trading with Russia, particularly in oil and after meetings with US senators, Rutte asked these nations to pressure Russian President Vladimir Putin into peace talks with Ukraine within 50 days. He echoed US President Donald Trump's threat of "biting tariffs" on Russian export buyers."If you live now in Beijing, or in Delhi, or you are the president of Brazil, you might want to take a look into this, because this might hit you very hard," Rutte said. This provocative statement directed at sovereign nations, including India, coming from the head of a military alliance, instantly raises questions about diplomatic overreach and to top it all, the sheer threat looms over countries like India, but strangely, the hammer never seems to fall on European nations, even though they're still buying Russian oil themselves. In the third year of the Russia-Ukraine War, Europe's fossil fuel imports from Moscow dropped by just 1% year-on-year. So, who exactly needs the lecture on responsible trading?This is why Rutte's bid to dictate India's trade policies is purely IS NOT A TRADE POLICEMANMark Rutte is the Secretary General of Nato, which is a military alliance focused on collective defence, and not a body governing global warning to India, a sovereign nation with no Nato affiliation, oversteps his comments followed Trump's announcement of new weapons for Ukraine. This hints that there is an alignment of Nato with the US policy, which is not at all has seen Brics as an anti-US bloc that is trying to bring in a currency that would undermine the American to forget, the US is a Nato member, and its biggest fact, Trump has been arm-twisting other Nato members, pushing them to spend more on defence Rutte's warning actually reveal his abject surrender to Trump's threats?Trade disputes and their resolution belong in fora like the World Trade Organisation (WTO), which, although, face their share of criticisms of bias, has the jurisdiction to address such issues. Nato's role is security, not economic coercion, and Rutte's foray into trade threats, parroting US President Donald Trump, confirms the NEEDS NO LECTURE ON PEACE, NON-ALIGNMENTadvertisementRutte's call for India to "make the phone call to Vladimir Putin" to push for peace is patronising and dismissive of India's diplomatic efforts. India has consistently advocated for peace, with Prime Minister Narendra Modi stating, "This is not the era of war", during global should India, China, or Brazil be the ones calling Putin? Did Trump's battering of Zelenskyy not work last time, or is the "daddy's" desperation to grab a Nobel Peace Prize getting more intense?India has balanced its strategic autonomy by abstaining from UN resolutions condemning Russia while engaging in dialogue with all parties. Rutte's allegation that India is not doing enough is not only unfounded but also insults New Delhi's proactive role in trying to foster dialogue between Russia and August 2024, Modi called up Russia's Putin, a day after he spoke to US President Joe Biden. The calls come close on the heels of his trip to a tone towards a strategic partner doesn't foster cooperation. It only alienates them. Though frankly, it's no surprise. The US and its administrations, including Trump's, have used this kind of language, or even harsher, for their own so-called "closest allies" CAN'T SANCTION. AND SHOULDN'T TRYThreatening India with sanctions for buying Russian oil, described as "slamming back in a massive way", is coercive, not diplomatic messaging. It aligns with Trump's aggressive tariff threats, but Nato certainly lacks the authority to impose increased imports of discounted Russian oil since 2022 is a pragmatic move driven by energy security, not an alarming alignment. Threatening a sovereign nation for prioritising its economic interests, especially when Western countries have historically done the same, frankly reeks of warning comes suspiciously close to the Brics summit, hosted by Brazil last week. The Brics' growing influence, expanding footprint, with discussions on alternatives to the petrodollar system gaining statement, following Trump's frustration with the Brics' "anti-American policies", appears less about Ukraine and more about countering the bloc's rising clout. This timing of the warning suggests it is also looking to tame the Brics' success, rather than a genuine call for IS STRAYING FROM ITS MANDATEThe use of Nato, a security alliance, to meddle in India's bilateral trade policies blurs critical lines between diplomacy and senators are pushing a bill for 500% tariffs on countries trading with Russia. It also hints towards a broader Western involvement in economic threats risks tarnishing its credibility as a security-focused entity, and casts Rutte as a mouthpiece for Washington's trade office rather than one of the security ENERGY POLICY IS NOBODY'S TO DICTATERutte's, in the warning, perhaps assumes India's energy policy is overly reliant on Russia. It is misleading. India has diversified its oil imports, sourcing from the Middle East, Africa, and the US. Meanwhile, it leverages discounted Russian oil to stabilise domestic prices. That's India's pragmatic approach of balancing economic interests with global India to abandon this strategy to align with Western demands ignores its energy security needs and the realities of a multipolar world is no longer unipolar, it's not even bipolar any more. We're clearly heading toward a multipolar warning to India is a clear violation of diplomatic boundaries. It is a condescending tone toward a sovereign nation, and a misuse of Nato's authority. It misreads India's independent global posture and its balanced approach to the Russia-Ukraine as it has for decades, acts responsibly. It prioritises its energy security and advocates for peace without succumbing to external pressure. As the External Affairs Minister S Jaishankar has, on multiple occasions, pointed out, the world must engage with India as an equal partner, not issue ultimatums through inappropriate channels. He said India did not need preachers, but statement wasn't just disrespectful. It was the wrong message, delivered in the wrong way, by absolutely the wrong person.- Ends
Mint
an hour ago
- Mint
Elon Musk's dubious deal for Tesla's investors is a hard sell
Tesla's investors, perhaps a little troubled by the American electric vehicle maker's sagging sales and profits, are to be offered a bonus: A chance to own a slice of an AI model that recently called itself 'MechaHitler." That is but one of the dubious benefits of Elon Musk's plan to have Tesla investors vote on the company investing in his artificial intelligence firm, xAI Holdings. xAI made headlines for all the wrong reasons last week when its chatbot Grok ran amok with anti-semitic replies to queries, followed swiftly by the resignation of Linda Yaccarino, chief executive of X, Musk's social media platform that was rolled into xAI earlier this year. If all that doesn't sound like the best time to pitch Tesla investors on buying in, you maybe weren't around for the SolarCity deal in 2016. Also Read: Musk's woes: Tesla hit refresh on its EVs but it hasn't worked That was when Musk urged Tesla shareholders to buy a faltering rooftop solar company in which he was chair, investor and creditor (his cousin was the CEO). At the time, Musk emphasized the centrality of solar power to Tesla's mission and unveiled a seemingly ingenious product, the solar roof. Today, the solar roof is notable chiefly for its rarity in the wild and Tesla stopped bothering to report solar installations some time ago. On the other hand, SolarCity was afforded a graceful exit and Musk was repaid the money he had lent to it. From that highly specific viewpoint, despite SolarCity's evident troubles, it was a great time for Tesla to buy. In this case, it appears Musk intends for Tesla to merely invest in xAI rather than buy it outright. There is scant insight into how xAI is doing financially, since it is privately held. We do know, however, that it issued $3 billion of secured debt, maturing in 2030, at the end of June with a coupon of 12.5%, currently yielding just over 12%. The rough 8 percentage points spread over Treasuries is in the same ballpark as triple-C rated doesn't suggest financiers are falling over themselves to fund xAI's substantial cash burn, put at over $1 billion a month by a Bloomberg News report in June. Nor does another report that xAI turned to Space X, another Musk vehicle, for $2 billion. Also Read: Mint Quick Edit: Elon Musk's new 'America Party' has spotted a gap in US politics On that basis, even a portion of Tesla's reported $37 billion cash balance would help Musk's AI project, in which Bloomberg estimates he holds a roughly one-third stake. Musk floated a figure of $5 billion when he polled X users on the idea last year. Confusingly, he posted about the plan for Tesla shareholders to vote on providing new funding on Sunday, two days after he rejected reports that xAI was raising fresh funds. Apart from obvious potential conflicts of interest, however, what would Tesla get out of it? I suppose you could argue that xAI's technology would help with realizing Tesla's autonomy and robotics ambitions. But with a full merger seemingly ruled out, for now at least, why would Tesla need to invest rather than just contract xAI's services? More importantly, Musk has been touting Tesla's in-house capabilities on these fronts for years, especially when it comes to robotaxis. The very idea that Tesla needs to kick several billion into a separate company in order to achieve its goals would undermine that narrative. Another possible reason for Tesla to do this would be to incentivize Musk to stay focused on the company. Recall that one of the arguments advanced last year for reinstating a massive options package for Musk that was struck down by a Delaware court was to ensure that he didn't take his best ideas elsewhere. Also Read: Jaspreet Bindra: Grok reflects the personality of its owner Elon Musk It was a breathtaking and egregious stance for the CEO and director of a major listed company, as well as the board, to take. Musk had founded xAI a year before, so he was already taking at least some ideas elsewhere. One thing to keep an eye on is whether Musk's ruling out of an acquisition ultimately holds. Based on current math, and using the reported $200 billion valuation target for xAI's next funding round, buying the company with stock would help him boost his current stake in Tesla from just under 13%, excluding options, to more than 16% in a combined entity, if all else is equal. In any case, Musk's pledges of focus aren't exactly rock solid. This is the man who in late April told investors he would step back significantly from his political work, but only to announce he was forming a whole new political party less than three months later. Regardless, the example of SolarCity, along with Tesla's outlandishly resilient stock, suggest any vote on investing in xAI would be likely to pass. When so much of Tesla's valuation rests on sheer investor faith in Musk, his wishes tend to be more like commands. ©Bloomberg The author is a Bloomberg Opinion columnist covering energy.
