SpaceX plans to launch Starship for the fifth time tomorrow, October 13th, during a 30-minute launch window that opens at 8AM ET. That’s after the Federal Aviation Administration (FAA) declared today that the company has “met all safety, environmental and other licensing requirements for the suborbital test flight,” reports Reuters.

The company will livestream the launch, starting about 35 minutes before liftoff. You can catch that on SpaceX’s website or its X account, as well as in the X TV app.

SpaceX’s goal for tomorrow’s flight is to return the Super Heavy booster to its launch site, which hasn’t happened before. It’s shooting for another successful splashdown of Starship in the Indian Ocean, which it pulled off for the first time in its fourth flight test in June.

The FAA’s clearance wasn’t anticipated until November, but the agency and its partner “conducted assessments more quickly than anticipated,” as CNBC reported yesterday. In addition to tomorrow’s planned launch, Reuters writes that the FAA “also approved the Starship 6 mission profile.”

We’ve just reviewed the Samsung Galaxy Buds 3 Pro and we’re very impressed: they deliver great-sounding audio quality, top-notch ANC and a premium, though derivative, design. But there’s always room for improvement, and there are often little irritations that only emerge after a product launch. Which is why Samsung has created a free software update just a few weeks after launch.

The update was spotted by Sammobile, which reported that according to Samsung’s rather terse release notes the updates “improve the stability of the Bluetooth connection and sound”. The update is also available for the non-Pro Buds 3.

How to get the Samsung Galaxy Buds 3 Pro software update

“No one gets to take from me and live. No one,” vows Sugar Goodson (Stephen Graham), a dangerous criminal kingpin, in the first teaser trailer for A Thousand Blows. The forthcoming drama is the latest TV offering from Peaky Blinders creator Steven Knight. Like his most successful and beloved TV effort, too, A Thousand Blows will both explore England’s history and immerse viewers in a criminal underworld full of tension, deceit, and violence.

Inspired by numerous real-life stories, the series is set in the East End of London in the 1880s. It follows two friends from Jamaica, Hezekiah (Malachi Kirby) and Alec (Francis Lovehall), as they find themselves fighting for their lives in East London’s criminal underground. Once there, Hezekiah and Alec not only cross paths with Mary Carr (Erin Doherty), the leader of an all-female gang, but also make an unfortunate enemy of Graham’s Sugar, whose criminal viciousness is reflected in his skill as an illegal boxer.

In addition to Graham, Kirby, Lovehall, and Doherty, the cast for A Thousand Blows also includes Ziggy Heath, Jason Tobin, James Nelson-Joyce, Tom Davis, and Ella Lily Hyland. The series recently premiered its first two episodes at this year’s London Film Festival. You can check out its debut trailer below.

Knight is one of the busiest creators working in Hollywood right now. One need only look at his slate of 2024 releases, which includes two high-profile projects he wrote in the Elisabeth Moss-led FX espionage thriller The Veil and Pablo Larraín’s new Angelina Jolie-starring Oscar contender Maria, for proof of that. Even by Knight’s standards, though, 2025 is shaping up to be a particularly noteworthy year.

Not only is A Thousand Blows expected to make its highly anticipated premiere next year but Netflix is also set to release Knight’s long-awaited feature film follow-up to Peaky Blinders. That project, titled The Immortal Man, began filming in late September with Heart of Stone director Tom Harper at its helm and Cillian Murphy officially back as Peaky Blinders anti-hero Tommy Shelby. The film, which is also set to star Graham, Rebecca Ferguson, Barry Keoghan, and Tim Roth opposite Murphy, boasts a screenplay penned by Knight.

There is, obviously, huge anticipation surrounding The Immortal Man‘s release. It looks like A Thousand Blows has the potential to be another breakout smash hit for Knight, though. At least, if its first teaser is any indication, it certainly looks like it’ll deliver similar vibes and thrills as Peaky Blinders. That possibility alone makes it one of next year’s most promising TV titles.

A Thousand Blows is set to premiere on Hulu sometime in 2025.

Apple is set to overtake Samsung as the world’s leading smartphone manufacturer in 2025, according to industry forecasts. The company’s success will be fueled by its strategic moves in both mature and emerging markets, along with the anticipated release of its budget-friendly iPhone SE 4.

Apple Expected to become largest smartphone maker in 2025

According to data from Counterpoint, Samsung currently holds a 3% lead over Apple in global market share as of the second quarter of this year. However, the company has historically performed well in the third and fourth quarters, benefitting from the annual iPhone release cycle.

Techinsights predicts that Apple will maintain its momentum through the first half of the year. If this forecast holds, the tech giant will dethrone Samsung, a position it has held for years.

Apple’s rise to becoming the largest smartphone maker in 2025 is driven by strong sales of older iPhone models. Emerging markets like India and China are leading this trend. The upcoming budget-friendly iPhone SE 4, priced at $499, will likely boost Apple’s market share further.

The innovations, including the rumored ultra-slim iPhone 17, will play a crucial role in this growth. The advanced Apple Intelligence platform will also impact this trend. These technologies are expected to speed up the iPhone replacement cycle. This strategy encourages users to upgrade to the latest models.

Apple’s focus on AI with its Apple Intelligence platform aims to deepen its integration with iOS and macOS.

In mature markets, Apple’s growth will benefit from favorable economic conditions. Exchange rates will also play a supportive role. These factors are expected to make Apple’s high-end devices more accessible. This accessibility will likely increase sales in regions where demand for premium smartphones remains strong.

Competitive Landscape: Samsung and Huawei’s Challenges

Huawei is also facing challenges as Apple gains ground. The company has been resilient, especially in recent years, but economic pressures and manufacturing constraints in China are expected to slow its momentum. These issues could limit Huawei’s ability to compete at the same level in the global market. Apple’s rise in 2025 may further widen the gap between Huawei and its rivals.

Microsoft is expanding Xbox Cloud Gaming’s streaming capabilities to support titles that aren’t on Xbox Game Pass, reports. The company has stated its plans to let users stream games from their own libraries numerous times over the last few years, going back to the announcement of , but the feature hasn’t yet materialized. Now, Microsoft is ready to start testing it among Xbox Insiders in November, with plans to ultimately support thousands of games, according to The Verge, which cites sources “familiar with Microsoft’s plans.”

Xbox players have long been asking for the option to stream games that they’ve purchased, rather than only those that are available on Xbox Game Pass. If the testing period goes as planned, more players could soon see the ability to do this. The report comes on the heels of Xbox president Sarah Bond’s announcement on Friday that Android users will be able to starting next month.

By now, large language models (LLMs) like ChatGPT and Claude have become an everyday word across the globe. Many people have started worrying that AI is coming for their jobs, so it is ironic to see almost all LLM-based systems flounder at a straightforward task: Counting the number of “r”s in the word “strawberry.” They are not exclusively failing at the alphabet “r”; other examples include counting “m”s in “mammal”, and “p”s in “hippopotamus.” In this article, I will break down the reason for these failures and provide a simple workaround.

LLMs are powerful AI systems trained on vast amounts of text to understand and generate human-like language. They excel at tasks like answering questions, translating languages, summarizing content and even generating creative writing by predicting and constructing coherent responses based on the input they receive. LLMs are designed to recognize patterns in text, which allows them to handle a wide range of language-related tasks with impressive accuracy.

Despite their prowess, failing at counting the number of “r”s in the word “strawberry” is a reminder that LLMs are not capable of “thinking” like humans. They do not process the information we feed them like a human would.

Almost all the current high performance LLMs are built on transformers. This deep learning architecture doesn’t directly ingest text as their input. They use a process called tokenization, which transforms the text into numerical representations, or tokens. Some tokens might be full words (like “monkey”), while others could be parts of a word (like “mon” and “key”). Each token is like a code that the model understands. By breaking everything down into tokens, the model can better predict the next token in a sentence. 

LLMs don’t memorize words; they try to understand how these tokens fit together in different ways, making them good at guessing what comes next. In the case of the word “hippopotamus,” the model might see the tokens of letters “hip,” “pop,” “o” and “tamus”, and not know that the word “hippopotamus” is made of the letters — “h”, “i”, “p”, “p”, “o”, “p”, “o”, “t”, “a”, “m”, “u”, “s”.

A model architecture that can directly look at individual letters without tokenizing them may potentially not have this problem, but for today’s transformer architectures, it is not computationally feasible.

Further, looking at how LLMs generate output text: They predict what the next word will be based on the previous input and output tokens. While this works for generating contextually aware human-like text, it is not suitable for simple tasks like counting letters. When asked to answer the number of “r”s in the word “strawberry”, LLMs are purely predicting the answer based on the structure of the input sentence.

Here’s a workaround

While LLMs might not be able to “think” or logically reason, they are adept at understanding structured text. A splendid example of structured text is computer code, of many many programming languages. If we ask ChatGPT to use Python to count the number of “r”s in “strawberry”, it will most likely get the correct answer. When there is a need for LLMs to do counting or any other task that may require logical reasoning or arithmetic computation, the broader software can be designed such that the prompts include asking the LLM to use a programming language to process the input query.

Conclusion

A simple letter counting experiment exposes a fundamental limitation of LLMs like ChatGPT and Claude. Despite their impressive capabilities in generating human-like text, writing code and answering any question thrown at them, these AI models cannot yet “think” like a human. The experiment shows the models for what they are, pattern matching predictive algorithms, and not “intelligence” capable of understanding or reasoning. However, having a prior knowledge of what type of prompts work well can alleviate the problem to some extent. As the integration of AI in our lives increases, recognizing its limitations is crucial for responsible usage and realistic expectations of these models.

 Chinmay Jog is a senior machine learning engineer at Pangiam.