Is there a realistic trade out there that works for Miami?
January 2025
Windows Is Expanding Its Start Menu Phone Integration to iPhones
Sharing files or accessing messages between a Windows PC and an iPhone can be an utter pain in the rear, mostly because, without third-party utilities, you can’t rely on AirDrop to do it. There’s the Phone Link app, but even that requires opening up a separate window—your PC won’t even be able to serve you iMessage notifications unless it’s open. Meanwhile, Android users currently have a smoother experience, able to access their connected phones’ battery levels, messages, and calls right from the Windows Start menu. Now, iPhone users are finally about to get the same.
In a blog post on its site, Microsoft announced that its Windows Insider beta testers are currently getting expanded Start menu phone integration, bringing features Android users have been enjoying since October to iPhone.
These include the ability to connect a phone to your PC right from the start menu, plus that battery status and message/call information (seemingly, you’ll finally be able to get your iPhone notifications on Windows without needing to open the Phone Link app first). The only thing that appears to be missing is the ability to browse your photos right from the start menu, but that doesn’t mean you won’t be able to access them.
That’s thanks to the new “Send files” button, an addition for both Android and iPhone. This will let you access your phone’s files without having to go through the Phone Link app, and it should make photo sharing much simpler, as doing so with an iPhone currently requires relying on either another test feature or going through the multi-step process to connect iCloud to the Windows Photos app.
To see iPhone integration in the Windows 11 Start menu, you’ll need to have an iPhone running iOS 16 or higher (meaning nothing older than an iPhone 8), plus the Phone Link app on both your iPhone and Windows 11 PC. You’ll also need to register for the free Windows Insider program, but most importantly, you’ll need to wait.
Microsoft says these features are “gradually rolling out to Windows Insiders in the Dev and Beta Channels,” so just being signed up might not be enough to access them at first. You can check if you have them by navigating to Settings > Personalization > Start.
NBA power rankings 2024-25: Thunder remain on top while Rockets vault up to second after dominating week
Is it time to consider Houston a legitimate title contender?
Tigers sign RHP Tommy Kahnle to 1-year deal worth $7.75 million
Kahnle, 35, went 0-2 with a 2.11 ERA and one save in 50 relief appearances with the AL champion New York Yankees last season.
RHP Ryne Stanek is reportedly rejoining the New York Mets on a 1-year deal
The Mets acquired Stanek from Seattle in exchange for minor league outfielder Rhylan Thomas at the trade deadline last year.
How I Learned That I Don’t Need a Carving Knife (and What I Use Instead)
We may earn a commission from links on this page.
If you’ve been shopping around for knives, you’ve probably noticed that there are tons of options available to you. Between boning knives, filet knives, bread knives and the like, it can start to feel like each ingredient needs its own special blade. Recently, I was wondering if a carving knife would be worth adding to my toolkit, so I got one to test out. It was sharp, light, and pretty to look at, but it turns out I didn’t need a carving knife at all.
I have been working with a lot of chicken lately—spatchcocked birds, butterflied chicken breasts, and deboned chicken thighs—and when you’re slicing a lot of meat you begin to wonder if there’s a “perfect” knife to do the job. I was carving meat, so a carving knife is what I needed. Right?
What is a carving knife?
The goal of a carving knife is to take care of long cuts or single-stroke cuts. The fewer times you have to lift the knife and continue the same cut, the better. It also might be your intention to use a carving knife to divide a roast into parts, like a chicken into drumsticks, wings, and breast slices. It can help you there too.
Carving knives vary slightly depending on the knife-maker or brand, but overall, you’ll see more similarities among them: The length (sometimes up to 12 inches) and a slightly curved and pointed blade. The sharp and thin blade ensures that the meat doesn’t tear, and the length makes it easier to slice large roasts.
A chef’s knife can do the trick
After trying out my carving knife for a few months, I was happy with its performance, but I wasn’t in love. I wanted to go back to using my chef’s knife for carving. It was just as sharp, I preferred the balance of its weight more, and the length wasn’t all that different, so I could still get through large roasts. What’s more, I really didn’t use the carving knife for anything else, so it was taking up valuable space on my magnet strip.
Sure, if you’re always smoking brisket or whole roasts are frequently a part of your family meal rotation, it might make sense for you to invest in a carving knife. If you do, I suggest one with a hollow edge like this one from Messermeister. The little divots in the sides provide a small air pocket between the metal and the meat to prevent suction.
For me and others who are only slicing big roasts a few times a season or once a year, a carving knife may not be a necessary expense if you have a chef’s knife you already like. I’ve had a Wüsthof chef’s knife for ages and it’s often the only knife I need to prepare a meal.
A good chef’s knife can do the chopping, scoring, and carving as long as you keep it sharp. If you’re not sure how to do that, do not fear: Here’s a helpful guide on taking care of your knives.
Inside the Lakers’ complex approach to NBA trade deadline decisions
The Lakers’ players would love to see a trade that bolsters the roster, but they acknowledge that most moves are complex and expensive.
AI: Alibaba Introduces Qwen 2.5-Max
By: Jeffrey Kondas, Technology Fellow, with Grok xAI
Alibaba has recently introduced several new AI models under its Qwen series, with the latest being Qwen 2.5-Max and an array of models named Qwen2.5-VL. Here’s a detailed overview based on the latest information and a comparison with top market AI:
Qwen 2.5-Max:
- Performance Claims:
- Alibaba claims that Qwen 2.5-Max outperforms OpenAI’s GPT-4o, DeepSeek’s V3, and Meta’s Llama 3.1-405B in various benchmark tests, including problem-solving, coding, and math benchmarks. It also performed on par with Anthropic’s Claude 3.5 Sonnet in some areas.
- Market Impact:
- Following the announcement, Alibaba’s stock saw significant gains, suggesting market approval of the new model’s capabilities.
- Tech Specifications:
- The specifics of its architecture or parameter count “are not fully disclosed”.
Qwen2.5-VL Series:
- Capabilities:
- These models are capable of parsing files, understanding videos, counting objects in images, and even controlling PCs and mobile devices. They can perform tasks similar to OpenAI’s Operator by interacting with software applications.
- Benchmarking:
- According to Alibaba, Qwen2.5-VL models have shown superior performance in video understanding, math, document analysis, and question-answering when compared to GPT-4o, Claude 3.5 Sonnet, and Gemini 2.0 Flash.
- Applications:
- They can analyze charts, extract data from invoices and forms, comprehend long videos, and recognize intellectual properties from movies and TV series. A notable feature is their ability to book flights or perform other tasks directly through app interfaces.
- Availability:
- The smaller models (Qwen2.5-VL-3B and Qwen2.5-VL-7B) are available under a permissive license, while the flagship model (Qwen2.5-VL-72B) operates under Alibaba’s custom license, which has specific commercial use restrictions for large enterprises.
Strategic Context:
- Competitive Landscape:
- Alibaba’s move is seen as a response to the rapid advancements and market disruptions caused by competitors like DeepSeek in China and international players like OpenAI and Meta. The timing of Qwen 2.5-Max’s release during the Lunar New Year underscores the competitive pressure.
- AI Price Wars:
- Alibaba, along with other Chinese tech companies, has been part of an AI price war, reducing costs to attract more users and developers, which is crucial for expanding their AI ecosystem.
- Open-Source Strategy:
- Alibaba has taken a hybrid approach, offering both proprietary and open-source models to cater to a broad audience and encourage wider adoption and contributions from the global AI community.
Qwen 2.5-Max Overview:
Features:
- Coding:
- Qwen 2.5-Max has shown competitive performance in coding tasks. According to benchmarks like HumanEval and MBPP, it scores 73.2 and 80.6 respectively, which suggests it’s on par or slightly better than models like DeepSeek V3 and significantly ahead of Llama 3.1-405B for coding tasks. This indicates its capability for not just code generation but also in understanding and resolving coding problems.
- Prompt and Response Token Limits:
- Qwen 2.5-Max supports a context length of up to 20 trillion tokens in training, one of the largest datasets known. However, the operational context window for user interaction is up to 131,072 tokens (128K for practical use).
- Character Count: Roughly, one token in English text equals about 3-4 characters. Thus, 131,072 tokens translate to approximately 393,216 to 524,288 characters.
Comparison to Market AI:
- Qwen 2.5-Max stands out with its large training dataset, which might contribute to its performance in coding and problem-solving tasks. Its token limit is competitive within the industry, though not the largest, indicating a balance between capability and efficiency.
- GPT-4o (OpenAI):
- Coding: Excels in both code generation and problem-solving, with scores around 69.2 on HumanEval but is noted for its versatility across languages and frameworks.
- Token Limits: Supports up to 128,000 tokens in its latest version, translating to around 384,000 to 512,000 characters.
- Source: OpenAI’s GPT-4o Capabilities
- Claude 3.5 Sonnet (Anthropic):
- Coding: Known for its accuracy and context understanding, scoring around 80% on code-related tasks but with a focus on ethical coding practices.
- Token Limits: Can handle up to 200,000 tokens, making it suitable for long-form content analysis, roughly 600,000 to 800,000 characters.
- Source: Anthropic’s Claude 3.5 Sonnet Announcement
- DeepSeek V3 (DeepSeek AI):
- Coding: Specifically designed to excel in coding tasks, it performs well in benchmarks like DS-FIM-Eval and DS-Arena-Code but lags behind Qwen 2.5-Max in some areas.
- Token Limits: Similar to Qwen 2.5-Max with a 128K token context window, translating to about 384,000 to 512,000 characters.
- Source: DeepSeek AI Blog on V3
- Gemini (Google):
- Coding: While versatile, it’s noted for its integration capabilities rather than being a top performer in coding benchmarks.
- Token Limits: Gemini’s largest model supports up to 2 million tokens, equating to around 6 million to 8 million characters, significantly outpacing others in context handling.
- Source: Google DeepMind Gemini
- Compared to competitors, Qwen 2.5-Max offers a strong coding capability, especially in benchmarks where it outperforms or matches high-profile models like GPT-4o and Claude 3.5 Sonnet. Its token limit is well-suited for most practical applications, but models like Gemini push the boundaries for handling extremely long contexts.
Disclaimer: Token to character conversion is approximate and can vary based on the text’s nature. The data here is based on the latest public information, which might evolve with new updates from these companies.
Check It:
Caitlin Clark declines invite to a 3-point shootout at NBA All-Star weekend
Caitlin Clark is going to wait until Indianapolis hosts the WNBA All-Star weekend in July to enter her first 3-point contest as a pro.
AI: Qwen 2.5-Max on the HumanEval and MBPP benchmarks
Here’s an expanded explanation of the performance of Qwen 2.5-Max on the HumanEval and MBPP benchmarks:
HumanEval Benchmark:
- Overview: HumanEval is a benchmark specifically designed to test the coding capabilities of AI models. It consists of a set of 164 Python programming problems, each with a function signature and a detailed description. The problems range from basic to moderately complex, covering various aspects of Python programming like data structures, algorithms, and basic syntax.
- Scoring: The score on HumanEval is typically calculated based on the percentage of problems where the AI model generates a correct solution that passes all given test cases. A score of 73.2 for Qwen 2.5-Max on HumanEval indicates that it successfully solved about 73.2% of these problems correctly. This is a high score, suggesting that Qwen 2.5-Max has a strong understanding of Python programming, can interpret requirements accurately, and can generate functional code.
- Implications:
- Code Generation: This score reflects Qwen 2.5-Max’s ability to generate code from scratch based on problem descriptions, demonstrating its proficiency in language understanding and code syntax.
- Problem Solving: It also shows the model’s capability in algorithmic thinking and problem decomposition, which are crucial for real-world software development.
MBPP (Mostly Basic Programming Problems) Benchmark:
- Overview: MBPP is another benchmark that tests coding ability, but it’s broader in scope, covering multiple languages like Python, Java, C++, and more. It includes 974 problems, ranging from simple to intermediate, which are designed to be solved by beginners or those with basic programming knowledge.
- Scoring: Similar to HumanEval, MBPP’s scoring is based on the pass rate of the generated solutions against provided test cases. Achieving a score of 80.6 on MBPP means that Qwen 2.5-Max successfully solved 80.6% of these problems across different languages or at least in Python if we assume the test was conducted in one language.
- Implications:
- Versatility: This high score indicates Qwen 2.5-Max’s versatility in handling different programming languages or its deep understanding when focused on one language like Python.
- Practical Coding: MBPP’s focus on basic to intermediate problems tests the AI’s ability in routine programming tasks, which are common in everyday development scenarios, thus suggesting its usefulness for developers in practical settings.
Combined Analysis:
- Comparative Performance: Qwen 2.5-Max’s scores on both benchmarks position it competitively among top-tier AI models. For context, these scores are higher than many contemporary models, though exact comparisons depend on the specific versions of other models tested at the same time under similar conditions.
- Use Case Fit: These results suggest that Qwen 2.5-Max could be particularly beneficial in environments where coding assistance, from ideation to code completion, is needed. Its performance in both benchmarks shows it can handle both the complexities of problem-solving (HumanEval) and the breadth of basic programming tasks (MBPP).
- Model’s Learning: The scores reflect the model’s training data quality and quantity, its architecture, and the effectiveness of its fine-tuning for coding tasks. The high performance might be attributed to exposure to a vast and diverse coding dataset during training or specialized fine-tuning for coding challenges.
- Future Considerations: While these benchmarks provide a snapshot of Qwen 2.5-Max’s capabilities, ongoing updates and the dynamic nature of AI development mean that these scores might improve or be challenged by newer models or versions of existing ones.
“By excelling in these benchmarks, Qwen 2.5-Max demonstrates its readiness to assist in coding tasks, potentially reducing development time and aiding in educational contexts by providing accurate solutions and explanations.”
- Reuters: Alibaba Unveils Qwen 2.5-Max AI Model
- Alibaba’s Qwen Blog: Qwen 2.5-Max: Features, DeepSeek V3 Comparison & More | DataCamp
- Qwen Official Site: Qwen2.5-Max: Exploring the Intelligence of Large-scale MoE Model
Please note that exact benchmark scores might not be directly quoted in these sources but are synthesized from various articles that discuss the capabilities and comparisons of Qwen 2.5-Max with other AI models. The specific scores of 73.2 on HumanEval and 80.6 on MBPP for Qwen 2.5-Max are based on the information provided in the query and might not be explicitly stated in these links but are consistent with the performance claims made by Alibaba and reported by tech news outlets.