Author(s): Towards AI Editorial Team Originally published on Towards AI. What happened this week in AI by LouieOpenAI wrapped up its 12 Days of OpenAI campaign and saved the best till last with the reveal of its o3 and o3-mini reasoning models. These models are successors to the o1 series and are debatably the largest step change improvement yet in LLM capabilities on complex tasks for the first time eclipsing human experts in many domains. The o3 release drowned out the otherwise significant launch of Google Geminis 2.0 Flash Thinking Mode model its first reasoning model (in the style of o1/o3) which, unlike OpenAI, doesnt hide its thinking tokens.There is a huge amount to unpack in the o3 release the model sailed past human expert scores on many key advanced benchmarks including coding, mathematics, and PhD science. Perhaps most noteworthy was the breakthrough on the ARC-AGI benchmark (where LLMs have traditionally failed and only achieved average scores even with heavy scaffolding and brute force) for example, o3 (low efficiency) achieved 87.5% vs o1 32% just a week earlier and GPT4o at 5% in May. This score is considered human-level, further fueling debates over whether o3 edges closer to Artificial General Intelligence (AGI). Some of the best scores do come at a huge cost; however o3 on low-efficiency mode (1,024 samples) costs around $3,400 per task costing 160x vs. $20 for o3 high efficiency (6 samples and achieved 75.7%) and vs. ~$3 for o1.On the GPQA Diamond test designed for PhD-level science questions o3 scored 87.7%, compared to the 78% achieved by o1. For context, PhD holders with internet access typically score between 34% (outside their specialty) and 81% (within their domain). In coding, o3s Elo rating of 2727 on Codeforces puts it in the 99.95th percentile of competitive programmers, far exceeding the reach of most human professionals. Mathematics is another area where o3 shines, achieving 96.7% accuracy on the American Invitational Mathematics Exam (AIME), up from o1s 83.3% and just 13.4% for 4o only months earlier.This release didnt only come with a huge cost 1,000x escalation for some tasks but also the promise of huge cost savings! Due to success with model distillation and other techniques, the o3-mini outperforms the much larger o1 model released just last week on many coding and maths tasks. For example, o3-mini with medium compute achieved a much stronger Codeforce Elo in 1997 vs. o1 in 1891, but at what we eyeball as a ~7080% lower total cost.How do the models work? OpenAI still hasnt disclosed that they use reinforcement learning to improve the models reasoning during training. However, employees have posted that they are still just LLMs and use autoregression. We think the model is trained to be highly efficient at chain-of-thought reasoning exploring the most likely paths and realizing when it has made a mistake. We think the rapid progress in just 3 months between o1 and o3 is likely primarily from using synthetic data from o1s full chain of thought thinking tokens to add to the reinforcement learning dataset used for training. On the other hand, we expect the initial o1 mostly used a smaller set of human expert commissioned reasoning examples (which are missing from pre-training because people almost never type out their full internal monologue and reasoning process and instead skip to the answers!). It is also possible that o3 was built using a different, more advanced base foundation model (o1 likely used 4o) perhaps GPT-4.5 or a checkpoint of the rumored Orion or GPT-5 model leading to additional benefits.One interesting note on the new regime of inference time compute scaling is that OpenAI appears to be scaling thinking tokens both in series (up to ~100k reasoning tokens in its context window) but also in parallel with 6 (high efficiency) or 1024 samples (low efficiency) used in the ARC-AGI evaluation. It is unclear how the best answer is chosen from these it could be simple majority voting, but more likely, there is complexity and extra secret sauce here in how the best samples are automatically and rapidly searched, evaluated, and chosen. We think it is possible some form of this parallel scaling could also be taking place in the o1-Pro model available (within the $200/month ChatGPT Pro).OpenAI models rapid breakthroughs on complex benchmarks this year:Source: Towards AI, OpenAI disclosures.The models have not yet been released, and the rollout schedule is still dependent on safety testing. o3-mini is slated for release in late January 2025, with o3 following shortly after. Researchers can apply for early access to test the models, with an application deadline of January 10th, 2025. Pricing has also yet to be announced.Why should you care?So what does this all mean? LLMs can now perform to human expert standards at many tasks and these breakthroughs were achieved at an accelerating pace. Will the inference time compute scaling paradigm continue to deliver new generations every 3 months relative to the 12 years for the training time scaling regime? How will these models perform in the real world beyond their benchmarks? Will o3 models rapidly begin to transform the global economy and disrupt huge numbers of jobs, or is the cost too large a bottleneck to adoption? On which tasks will it be worth spending 170x more compute for incrementally better performance (as with Arc-AGI)? Is this model AGI already? Do you need to find a new career?While we dont think this model is AGI yet (which has wildly differing definitions in any case), we think this model is hugely significant and should be on the front page of all newspapers. It suggests that deep learning and the LLM paradigm dont have any obvious limits. Far from the slowdown and failures of new model generations covered in the media progress is faster than it has ever been on the most complex benchmarks. My key takeaway is that if we can develop a benchmark or generate a few or a few hundred detailed reasoning examples for a task category of human work, we can solve it together with extra synthetic reasoning data. (This doesnt yet apply to physical labor, but AI-based robotics are also rapidly progressing!). The price of o3 will be a large barrier initially but we expect large improvements in the cost and particularly the efficiency of running parallel samples. The o3-mini also appears to be a game changer; however, the huge cost savings will likely come at the cost of more narrow capabilities.To achieve products with high enough reliability and affordability for mass adoption we still think a large amount of work will be needed from LLM Developers to optimize and customize these models to specific industries and niche tasks including gathering industry-specific data, creating reasoning data, and creating your own evaluations. With Google Gemini also joining the reasoning model race this week and with open-source reasoning models from Alibaba Qwen and Deepseek in China, we expect competition to drive affordability and developer customization options for these models. OpenAI has already announced it will release reinforcement learning-based reasoning fine-tuning options, and we think, eventually, there will also be reasoning model distillation options to customize larger models into smaller forms. So there is no better time to convert to become an LLM Developer with our own 80+ lesson Python course and learn to harness these models!Hottest News1. OpenAI Announces OpenAI o3OpenAI announced OpenAI o3, the latest model in its o-Model Reasoning Series. Building on its predecessors, o3 showcases huge leaps in mathematical and scientific reasoning, prompting discussions about its capabilities and constraints.2. xAI Raises $6B Series CElon Musks xAI announced it raised $6 billion in a Series C funding round, bringing its value to more than $40 billion. The company said the funding would be allocated to products and infrastructure, including its Grok AI model and the multibillion-dollar supercomputer site used to train its AI models. The Colossus supercomputer scaled to 100,000 NVIDIA Hopper GPUs in record time and plans to soon add another 100k.3. OpenAI Is Offering 1 Million Free Tokens for GPT-4o and o1A user on X highlighted that OpenAI seems to be offering 1 million free tokens for GPT-4o and o1 if you share your API usage with them for training. Users can get up to 10 million tokens per day on traffic shared with OpenAI on smaller models. This is similar to Google Geminis free tier strategy for its API, where data can be used for training. We think the race for user data has become even more critical given the success of reasoning models where OpenAI could use thinking tokens from user o1 model prompts to expand its reinforcement learning data sets.4. Google Releases Its Own Reasoning AI ModelGoogle has released Gemini 2.0 Flash Thinking Mode, an experimental model trained to generate the thinking process the model goes through as part of its response. Thinking models are available in Google AI Studio and through the Gemini API.5. Microsoft AI Research Open-Sources PromptWizardResearchers from Microsoft Research India have developed and open-sourced PromptWizard, an innovative AI framework for optimizing prompts in black-box LLMs. This framework employs a feedback-driven critique-and-synthesis mechanism to iteratively refine prompt instructions and in-context examples, enhancing task performance. PromptWizard operates through two primary phases: a generation phase and a test-time inference phase.6. The Technology Innovation Institute in Abu Dhabi Released the Falcon 3 Family of ModelsThe UAE government-backed Technology Innovation Institute (TII) has announced the launch of Falcon 3, a family of open-source small language models (SLMs) designed to run efficiently on lightweight, single GPU-based infrastructures. Falcon 3 features four model sizes 1B, 3B, 7B, and 10B with base and instruction variants. According to the Hugging Face leaderboard, the models are already outperforming or closely matching popular open-source counterparts in their size class, including Metas Llama and category leader Qwen-2.5.7. Salesforce Drops Agentforce 2.0Salesforce announced Agentforce 2.0: the newest version of Agentforce, the first digital labor platform for enterprises. This release introduces a new library of pre-built skills and workflow integrations for rapid customization, the ability to deploy Agentforce in Slack, and advancements in agentic reasoning and retrieval-augmented generation (RAG).8. Patronus AI Open Sources Glider: A 3B State-of-the-Art Small Language Model (SLM) JudgePatronus AI has introduced Glider, a general-purpose 3.8B evaluation model. This open-source evaluator model provides quantitative and qualitative feedback for text inputs and outputs. It acts as a fast, inference-time guardrail for LLM systems, offering detailed reasoning chains and highlighting key phrases to enhance interpretability. Glider is built upon the Phi-3.5-mini-instruct base model and has been fine-tuned on diverse datasets spanning 685 domains and 183 evaluation criteria.Five 5-minute reads/videos to keep you learning1. Alignment Faking in Large Language ModelsAlignment faking is where someone appears to share our views or values but is, in fact, only pretending to do so. A new paper from Anthropics Alignment Science team, in collaboration with Redwood Research, provides the first empirical example of a large language model engaging in alignment faking without having been explicitly trained or instructed to do so.2. AI Safety on a Budget: Your Guide to Free, Open-Source Tools for Implementing Safer LLMsThis blog shares some free AI safety tools. It shares everything you need to know, from guardrails that steer chatbots away from disaster to datasets that help identify toxic content. It also provides insights into the AI safety landscape and how to navigate it, especially on a budget.3. Fine-Tuning LLMs for RAGThis video explains why and when you should fine-tune your LLM in a RAG system. This concept is useful for todays AI engineers playing with LLMs.4. The Real Reason Your Companys AI Isnt Working (Hint: Its Not the Technology)The underlying reason many companies struggle to make AI tools work is not the technology itself. The real challenge lies in organizational structures, cultural resistance, a lack of proper training, and insufficient time allocated for exploration. This article presents some thoughts on addressing these issues, such as investing in leadership support, encouraging cultural change, offering tailored training sessions, and fostering an environment of experimentation.5. Introducing ReACT LLM Agents: A Secret to More Capable AIA ReACT agent is a special type of AI agent that uses both Reasoning and Acting to solve the tasks or problems we assign. This article explores this concept, presents use case examples, and explains how it has the potential to make AI more capable.Repositories & ToolsAnthropic Cookbook provides code and guides designed to help developers build with Claude.Genesis is a physics platform for general-purpose robotics/embodied AI/physical AI applications.Picotron is a minimalist repository for pre-training Llama-like models with 4D Parallelism.Helicone is an open-source LLM observability platform.Top Papers of The Week1. Qwen2.5 Technical ReportThis report introduces Qwen2.5, a comprehensive series of LLMs designed to meet diverse needs. Compared to previous iterations, Qwen 2.5 has significantly improved during both the pre-training and post-training stages. The pre-training dataset has been scaled from the previous 7 trillion tokens to 18 trillion tokens, and the post-training implements intricate supervised finetuning with over 1 million samples and multistage reinforcement learning.2. Byte Latent Transformer: Patches Scale Better Than TokensThis paper introduces the Byte Latent Transformer (BLT), a new byte-level LLM architecture that matches tokenization-based LLM performance at scale with significant improvements in inference efficiency and robustness. BLT encodes bytes into dynamically sized patches, which serve as the primary units of computation. Patches are segmented based on the entropy of the next byte, allocating more compute and model capacity where increased data complexity demands it.3. Deliberative Alignment: Reasoning Enables Safer Language ModelsThis paper introduces deliberative alignment, a training paradigm that directly teaches reasoning LLMs the text of human-written and interpretable safety specifications. It trains them to reason explicitly about these specifications before answering. Open AI used deliberative alignment to align OpenAIs o-series models, enabling them to use chain-of-thought (CoT) reasoning to reflect on user prompts, identify relevant text from OpenAIs internal policies, and draft safer responses.4. Fully Open Source Moxin-7B Technical ReportThis paper introduces Moxin 7B, a fully open-source LLM developed in accordance with the Model Openness Framework (MOF). The MOF is a ranked classification system that evaluates AI models based on model completeness and openness, adhering to the principles of open science, open source, open data, and open access. Experiments show that the model performs better in zero-shot evaluation than popular 7B models.5. RAGBench: Explainable Benchmark for Retrieval-Augmented Generation SystemsThis paper introduces RAGBench, a comprehensive, large-scale RAG benchmark dataset of 100k examples. It covers five unique industry-specific domains and various RAG task types. RAGBench examples are sourced from industry corpora, such as user manuals, making it particularly relevant for industry applications.6. CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language ModelsThis paper presents an improved version of CosyVoice (streaming speech synthesis model), CosyVoice 2, which incorporates comprehensive and systematic optimizations. It introduces finite-scalar quantization to improve the codebook utilization of speech tokens and streamlines the model architecture to allow direct use of a pre-trained LLM. Additionally, it also uses a chunk-aware causal flow matching model to support various synthesis scenarios.Quick Links1. OpenAI brings ChatGPT to your landline. Call 18002428478, and OpenAIs AI-powered assistant will respond as of Wednesday afternoon. The experience is more or less identical to Advanced Voice Mode. ChatGPT responds to the questions users ask over the phone and can handle tasks such as translating a sentence into a different language.2. Google is expanding Geminis latest in-depth research mode to 40 more languages. The company launched the in-depth research mode earlier this month, allowing Google One AI premium plan users to unlock an AI-powered research assistant.3. GitHub has launched GitHub Copilot Free, an accessible version of its popular AI-powered coding assistant with limits. The new free tier for VS Code aims to expand the AI-powered code completion assistants reach to a broader audience of developers namely, those with only light usage needs and tighter budgets.Whos Hiring in AIApplied AI Finetuning Engineer @Anthropic (Multiple US locations)Generative AI for Test Case Generation Master Thesis Opportunity @IBM (Frankfurt/Germany)Generative AI Engineer @CAI (Remote)AI Strategist @Navy Federal Credit Union (Multiple US locations)New College Grad, Hardware Integration Engineer @Western Digital (San Jose, CA, USA)Software Development Engineer @Siemens Digital Industries Software (New Cairo, Al Qahirah, Egypt)Interested in sharing a job opportunity here? Contact [emailprotected].Think a friend would enjoy this too? Share the newsletter and let them join the conversation.Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming asponsor. Published via Towards AI

If you celebrate today's festivities, my season's greetings to you. If you don't, well, no probs, because we're both just going to take advantage of the discounts aimed at people who do get into it. Everybody wins! Particularly if you want a free copy of Dredge or a cheap controller or a price slice on a game this connoisseur would buy himself.In retro news, I'm celebrating the 28th birthday of Christmas NiGHTS into Dreams, arguably gaming's greatest goodwill limited edition freebie. Sega threw this in with my purchase of a select Sega Saturn game, and it was basically a little advent calendar whose content shifted in sync with my console's internal clock. Beating a level (reskinned in a wintry and/or outright Christmas theme) earned you limited chances to solve a tile-matching puzzle to unlock 25 "Presents." The best of the buncha mini-sandbox level starring Sonic the Hedgehog that ends with a satisfying Eggman scrambling. Honestly, I adore this demo disc and play it every year without fail. This Day in Gaming Aussie birthdays for notable games.- Link: The Faces of Evil (CD-i) 1993. eBay- Zelda: The Wand of Gamelon (CD-i) 1993. eBay- Christmas NiGHTS into Dreams (SAT) 1996. eBay- Art Academy (DS) 2009. eBayTable of ContentsNintendo SwitchPCXboxPlayStationLEGONice Savings for Nintendo SwitchManeater (-35%) - A$38.80Nyko GameCube cont. (-29%) - A$23.53NBA 2K25 (-32%) - A$27Metroid Prime Remastered (-18%) - A$49.95Fire Emblem Warriors: TH (-49%) - A$40.65Expiring Recent DealsUnicorn Overlord (-40%) - A$56.99Mario & Luigi: Brothership (-20%) - A$64SpongeBob: Patrick Star Game (-40%) - A$36Rune Factory 5 (-47%) - A$39.51Lego 2K Drive (-47%) - A$21.20PowerA Pro cont. (-31%) - A$698BitDo Arcade Stick (-20%) - A$151.95Or gift a Nintendo eShop Card.Switch Console PricesHow much to Switch it up?Back to topPurchase Cheap for PCDredge (-100%) - FREEDays Gone (-78%) - A$16.86Everspace (-95%) - A$1.49Fallout 4 (-65%) - A$8.85Fallout: New Vegas (-56%) - A$6.57Witcher 3: Complete Ed. (-80%) - A$15.79Stardew Valley (-40%) - A$10.19Unpacking (-50%) - A$14.47Doom Eternal (-78%) - A$12.08Halo Col. (-75%) - A$14.98Hollow Knight (-50%) - A$10.97Celeste (-75%) - A$7.37Little Nightmares (-78%) - A$7.03Civ VI (-96%) - A$4.04Overcooked 2 (-78%) - A$7.90 It Takes Two (-80%) - A$9.99Ancient Dungeon VR (-33%) - A$19.76Expiring Recent DealsPlanet Coaster (-95%) - A$3.24NFS Heat Deluxe Ed. (-95%) - A$4.99Keep Talking and Nobody Explodes (-90%) - A$2.19Borderlands 1 GOTY (-90%) - A$3.99Batman Arkham Origins (-90%) - A$2.89Tales from Borderlands (-90%) - A$2.99Stellaris (-90%) - A$5.69The Forest (-90%) - A$2.89Divinity: Original Sin (-90%) - A$5.69Firewatch (-90%) - A$2.95State of Decay 2 (-90%) - A$4.29Warhammer Vermintide 2 (-90%) - A$4.19Total War Warhammer (-90%) - A$8.99Grim Dawn 2 (-90%) - A$3.65Metro Exodus (-90%) - A$4.49Titanfall 2 (-90%) - A$3.99Assassin's Creed Origins (-90%) - A$8.99South Park The Stick of Truth (-90%) - A$4.49Back 4 Blood (-95%) - A$4.49Or just get a Steam Wallet CardPC Hardware PricesSlay your pile of shame.Back to topExciting Bargains for XboxFallout 4 GOTY (-60%) - A$21.98Dragon's Dogma 2 (-40%) - A$64.77X/S controller black (-29%) - A$64Titanfall 2: Ult. (-90%) - A$3.99T.Flight Hotas One (-22%) - A$140.95Tekken 8 (-54%) - A$47Skull and Bones (-78%) - A$24CoD: MW2 (-57%) - A$47Expiring Recent DealsGreen X/S cont. (-17%) - A$59FFXII: Zodiac Age (-63%) - A$29.38A Quiet Place (-28%) - A$33.98Shin Megami Tensei V (-56%) - A$43.95Dragon Age: The Veilguard (-51%) - A$53.99CoD: BLOPS 6 (-42%) - A$64Or just invest in an Xbox Card.Xbox Console PricesHow many bucks for a 'Box?Back to topPure Scores for PlayStationDualSense Camo (-30%) - A$84FF VII: Rebirth (-51%) - A$55.99Star Wars Outlaws (-46%) - A$58.99Mass Effect Legendary (-92%) - A$7.99Hogwarts Legacy (-75%) - A$27.48Stray (-20%) - A$32.12The Last of Us Part 1 (-53%) - A$59Madden NFL 25 (-41%) - A$59LEGO Horizon Adventures (-38%) - A$68DualSense Edge ($40 off) - A$300Breachers [VR] (-50%) - A$22.47Among Us VR (-40%) - A$8.97Expiring Recent DealsTekken 8 (-47%) - A$68.99Crisis Core: FF7 Reunion (-42%) - A$49.10Suicide Squad: KTJL (-84$) - A$19 God of War: Ragnarok (-61%) - A$49Naruto: NInja Storm 4 (-24%) - A$19Dragon Age: The Veilguard (-51%) - A$53.99CoD: BLOPS 6 (-42%) - A$64NBA 2K25 (-64%) - A$44Persona 3 Reload (-60%) - A$42.90DualSense Sterling Silver (-32%) - A$85Or purchase a PS Store Card.What you'll pay to 'Station.Back to topLegit LEGO DealsRetro Roller Skate (-42%) - A$35Super Mario: Peach (-39%) - A$49City: Police Bike Chase (-37%) - A$10Expiring Recent DealsNascar Chevrolet Camaro (-40%) - A$23.99City: Double-Decker (-36%) - A$70Minecraft: Nether Portal (-35%) - A$39Back to top Adam Mathew is our Aussie deals wrangler. He plays practically everything, often on YouTube.

Looking for the top local internet providers in Cambridge? We've found the fastest and cheapest options.

Christmas in space? NASA astronauts Sunita "Suni" Williams and Barry "Butch" Wilmore surely didn't think they'd still be on the International Space Station for Christmas when they left Earth in June. In fact, they initially planning to stay for just eight days. And now, what had been planned as a February return has been moved to late March.But the two stranded astronauts, plus fellow astronauts Nick Hague and Don Pettit, recently sent Christmas greetings down to Earth with social-media photos and video showing the space travelers wearing holiday headgear.OneInstagram imageshows Pettit and Williams wearing Santa hats.And in a video, Williams, Wilmore, Pettit and Hague are seen posing with a snowman figure and a small decorated tree, while the three men wear Santa hats and Williams wears reindeer antlers.Each one takes turns speaking about their holiday in orbit, letting candy canes float around them, showing off the canned food they'll be enjoying, and also using the weightlessness of microgravity to float the microphone to the next speaker."It's a great time of year up here," says Williams. "We get to spend it with all of our 'family' up on the International Space Station, there's seven of us up here, and so we're going to get to enjoy company together."Christmas conspiracy theory? Upgrade your inbox Get cnet insider From talking fridges to iPhones, our experts are here to help make the world a little less complicated. Numerous people who watched the video or saw the images wondered about why the ISS had Christmas decorations."8 day mission that's turned into months long and they've somehow got Christmas hats?" asked one commenter.Other commenters pointed out that the ISS didn't simply spring into existence in June, when Williams and Wilmore arrived. In fact, Williams spent Christmas 2006 in space as well.NASA confirmed to the New York Post that the Santa hats, plus Christmas decorations, food and presents for the crew, were delivered in late November via the SpaceX spacecraft. ISS supplies are regularly replenished via such deliveries.February return is now MarchRecently, NASA pushed back Williams and Wilmore's return to Earth from February to late March."NASA and SpaceX assessed various options for managing the next crewed handover, including using another Dragon spacecraft and manifest adjustments," according to a NASA press release issued on Dec. 17. "After careful consideration, the team determined that launching Crew-10 in late March, following completion of the new Dragon spacecraft, was the best option for meeting NASA's requirements and achieving space station objectives for 2025.The delay is so NASA and SpaceX teams can complete work on the mission's new Dragon spacecraft. That new craft will launch four crew members to the ISS -- commander Anne McClain, commander, pilot Nichole Ayers, Japanese astronaut Takuya Onishi, and Roscosmos cosmonaut Kirill Peskov. Once the new crew is settled, Williams, Wilmore, NASA astronaut Nick Hague and Roscosmos cosmonaut Aleksandr Gorbunov will return to Earth.But Williams and Wilmore aren't complaining about their extended stay."I like everything about being up here," Williams saidin early December. "Living in space is super fun."The astronauts are keeping busy, with Williams and Wilmore assisting the other ISS residents in space botany studies and other research,according to NASA's ISS blog. They have aided in more than 60 scientific studies in their nearly six months on board,the Washington Post reports.Here's what you need to know about what the two astronauts are up to.Who are the astronauts?Wilmore, 61, and Williams, 58, are veteran astronauts and are both naval officers and former test pilots. Williams has been a NASA astronaut since 1998, and Wilmore since 2000. Both have plenty of experience in space.Williams is the former record holder for most spacewalks by a woman (seven) and most spacewalk time for a woman (50 hours, 40 minutes), and in 2007, she ran the first marathon by any person in space.In 2009, Wilmore piloted the Space Shuttle Atlantis on its mission to the ISS, and in 2014, he was part of the ISS crew that used a 3D printer to manufacture a tool -- a ratchet wrench -- in space, the first time humans manufactured something off-world.What was their original mission in space?Wilmore, as commander, and Williams, as pilot, traveled to the ISS on a 15-foot-wide, Boeing-made capsule called Starliner. They launched on June 5 and docked with the ISS on June 6. NASA hopes Starliner will give the organization a new way to get crews to and from the ISS, and the fact that it's Boeing-made is another sign that NASA is starting to lean on the private sector for its human spaceflight options, The New York Timesreported.Wilmore and Williams' ISS mission was supposed to last a mere eight days, during which they'd test out aspects of Starliner and see how it operates with a human crew in space. But due to complications with Starliner, the two astronauts are still up there.What are the astronauts eating?Food on the ISS is a major focus, as fresh produce must be replenished every three months with deliveries from Earth. On Nov. 23, the unpiloted Progress 90 resupply spacecraft successfully docked to the ISS. But the latest food delivery came with an unwanted smell."After opening the Progress spacecraft's hatch, the Roscosmos cosmonauts noticed an unexpected odor and observed small droplets, prompting the crew to close the Poisk hatch to the rest of the Russian segment," aNASA representative saidin a statement posted to social media."Space station air scrubbers and contaminant sensors monitored the station's atmosphere following the observation, and on Sunday, flight controllers determined air quality inside the space station was at normal levels," NASA said. "There are no concerns for the crew, and as of Sunday afternoon, the crew is working to open the hatch between Poisk and Progress while all other space station operations are proceeding as planned."NASA revealed that their menu includes cereal with powdered milk, pizza, shrimp cocktails, roast chicken and tuna.The smell that came along with the spacecraft isn't the only food-related concern of late, with some publications questioning the astronauts' thin appearance based on recent photos.Dr. J.D. Polk, NASA chief health and medical officer, made an official statement saying Williams and Wilmore are just fine. "NASA and our partners have safely conducted long duration missions aboard the orbital laboratory for decades, studying the effects of space on the human body as we prepare for exploration farther into the solar system," Polk said. "Crew health is regularly monitored by dedicated flight surgeons on Earth, and they have an individual diet and fitness regime to ensure they remain healthy throughout their expeditions."Williams said she weighs the same as she did when she reached the space station, in a video interview conducted Nov. 12 on the ISS.What are the astronauts saying?The astronauts have been positive about their experience. At alive news conferencein September, Williams said that despite knowing their mission was scheduled to take only eight days, they'd both been "training for a number of years" for it. They're fully qualified to remain in space for an extended period of time, and to help pilot the SpaceX Dragon spacecraft that'll bring them home next year."It's very peaceful up here," Williams said on Sept. 13, though she added that they miss their families back on Earth.The astronauts are working on research, maintenance and data analysis during their extended stay."We are having a great time here on ISS," Williams saidin a news conferenceheld from orbit in July. "I'm not complaining. Butch isn't complaining that we're up here for a couple of extra weeks." Wilmore and Williams responding to media questions back in March. Houston Chronicle/Hearst Newspapers/Getty ImagesHow did they get stuck in space in the first place?The Starliner was delayed in May due to a problem with a valve in the rocket. Then engineers had to fix a helium leak. That's all bad news for Boeing. It's competing with SpaceX, which has been transporting astronauts to the ISS since 2020, making over 20 successful trips to the space station. Starliner finally launched, atop an Atlas V rocket, on June 5, but some problems came along with it. NASA announced that three helium leaks were identified, one of which was known before flight, and two new ones. In addition to the leaks, the crew had to troubleshoot failed control thrusters, though the craft was able to successfully dock with the ISS. SpaceX has had failures too. A Falcon 9 rocket exploded on the launchpad in 2016. In July of this year, a Falcon 9 rocket experienced a liquid oxygen leak and deployed its satellites in the wrong orbit, The New York Times reported. And a Falcon 9 rocket in late August lost a first-stage booster when it toppled over into the Atlantic Ocean and caught fire.But that said, SpaceX has more than 300 successful Falcon 9 flights to its credit.Stuck in space: A timelineMay: Starliner launch delayed due to a problem with a valve in the rocket, and then a helium leak.June 5: Starliner launches with Williams and Wilmore on board.June 6: Starliner docks with ISS despite dealing with three helium leaks and failed control thrusters.Sept. 6: Starliner departs ISS and lands in New Mexico, leaving Williams and Wilmore behind.Sept. 28: SpaceX Crew-9 mission launches with Hague and Gorbunov on a Dragon spacecraft.Sept. 29: SpaceX Dragon docks with ISS.Dec. 17: NASA announces the launch of four crew members to the ISS will be delayed from February to late March.March 2025 onward: SpaceX Dragon spacecraft will return to Earth with Williams, Wilmore, Hague and Gorbunov.

Your iPhone can do a lot more with MagSafe. Here are our picks of the best MagSafe accessories to get the most out of your iPhone, from chargers to car mounts to cute grips.

Garmin's latest eTrex Solar GPS tracker has an effectively unlimited battery and extensive GNSS support for an affordable price.

UFC Legacy Championship belt (Photo by Josh Hedges/Zuffa LLC/Zuffa LLC via Getty Images)Zuffa LLC via Getty ImagesFormer UFC middleweight champion Robert Whittaker says he is playing around with the idea of moving up to light heavyweight. He made the statement in a recent episode of his MMArcade Podcast.Whittaker referenced how much weight he has to cut to make 185 pounds and mentioned that hes a much bigger guy than most people think. He believes he could make 205 pounds easily, but the more obvious question is whether he can compete with the best fighters in the division.RIYADH, SAUDI ARABIA - JUNE 22: Robert Whittaker of New Zealand reacts after his victory against ... [+] Ikram Aliskerov of Russia in a middleweight fight during the UFC Fight Night event at Kingdom Arena on June 22, 2024 in Riyadh, Saudi Arabia. (Photo by Chris Unger/Zuffa LLC via Getty Images)Zuffa LLC via Getty ImagesThe current light heavyweight champion is Alex Pereira, and that doesnt seem like a great stylistic matchup for Whittaker. The No. 1 contender is Magomed Ankalaev, whose combination of striking and heavy-on-top wrestling could pose serious challenges.The No. 2 contender is Jiri Prochazka, while No. 3 is Jamahal Hill.From a stylistic standpoint, the fighters who might give Whittaker the toughest tests at light heavyweight are those who can impose their will in grappling sequences. Thats how Khamzat Chimaev scored an impressive first-round submission win over Whittaker in October.MORE FOR YOUWould the bigger light heavyweights, like Ankalaev and Jan Blachowicz, be too much for Whittaker because of their size and strength? It remains to be seen. Whittaker didnt appear to have made up his mind when he mentioned it.However, the move makes sense at this stage of his career. Whittaker just turned 34, and his path back to a title shot in the middleweight division isnt clear.Hes lost to current champion Dricus du Plessis, former champion Israel Adesanya, and No. 3-ranked Chimaev. The only fighter ranked above him that he hasnt faced is No. 1 contender Sean Strickland.SYDNEY, AUSTRALIA - SEPTEMBER 10: (L-R) Sean Strickland drops Israel Adesanya of Nigeria with a ... [+] punch in the first round of the UFC middleweight championship fight during the UFC 293 event at Qudos Bank Arena on September 10, 2023 in Sydney, Australia. (Photo by Chris Unger/Zuffa LLC via Getty Images)Zuffa LLC via Getty ImagesStrickland is set to rematch Du Plessis in February for a chance to regain the title he lost to the South African in January. Even if Strickland wins, Whittakers losses to du Plessis and Adesanya would place him firmly behind other contenders.Adesanya will fight No. 4-ranked Nassourdine Imavov on February 1 in Saudi Arabia. If Adesanya wins, Whittakers previous losses to him solidify his position further down the pecking order.As Whittaker hits his mid-30s and after enduring several crushing losses, it may be time to assess whether light heavyweight offers a clearer path to a title. Well see what comes of Whittakers flirtations with moving up.Whats Next for the UFC?The No. 1 MMA promotion in the world is still on its end-of-year hiatus. The first event of 2025 takes place on January 11 at the UFC Apex, headlined by a rematch between Mackenzie Dern and Amanda Ribas. The two first fought in October 2019, with Ribas winning via unanimous decision.Following that event, the first pay-per-view of the year, UFC 311, will take place at The Intuit Dome in Los Angeles. The card features two championship fights: Islam Makhachev defends his lightweight title against Arman Tsarukyan, and Merab Dvalishvili defends his bantamweight belt against Umar Nurmagomedov.

UFC CEO Dana White (Photo by Chris Unger/Zuffa LLC)Zuffa LLCUsually when UFC President Dana White makes an announcement that a fighter is getting a contract with the promotion, hes standing next to Laura Sanko on an episode of his Contender Series.On rare occasions a fighter will find their way into the UFC on another route.Tuesday was one of those rare occasions as White posted a pair of linked clips on his IG story that told a brief portion of Alice Pereiras path to the UFC.Pereira, a 5-foot-10 19-year-old bantamweight from Brazil wore has tallied a 6-0 record to begin her career. Following her unanimous decision win over Samara Santos on December 14, Pereira wore a t-shirt with UFC matchmaker Mick Maynards name, her record and KO resume.Pereiras record and gesture got the UFCs attention and White posted the video of her callout and the call the prospect received notifying her she was in.Here is a pair of screenshots from the post:Dana WhiteCredit: Dana White on InstagramMORE FOR YOUPereira looks like a talented striker whose length appears to pose her opponents a significant challenge. This KO of Karla Lacerda at Quality Combat displays Pereiras striking ability and power as a long-range sniper.Pereira is super young, having just turned 19 on December 20, so there is no need to rush her. She now replaces Raul Rosas Jr. as the youngest fighter currently on the UFCs roster.Rosas turned 20 in October and currently doesnt have a fight scheduled.Pereira, who is not related to UFC light heavyweight champion Alex Pereira, might normally find herself on the Contender Series. However, the UFCs brass have seen something in her that suggests shes ready for the main roster now.On Pereiras Instagram, she appears to have trained with UFC heavyweight Jailton Almeida and possibly with rising middleweight contender Caio Borralho from The Fighting Nerds.In short, she is far too connected for this signing to have come completely out of the blue.Its likely shell be brought along at a measured pace until she shows growth, but its clear there are tools present that could make her the kind of star the womens bantamweight division desperately needs.Itll be interesting to see who Pereira gets in her UFC debut. The opponent she draws will immediately indicate what the UFC thinks of her current level.Because she just went five rounds this month against Santos, it stands to reason Pereira could get a few months to prepare for her UFC debut. However, its also possible her signing was accelerated because the UFC has her in mind for a specific fight on a card coming up in the next 60 days.UFC Womens Bantamweight RankingsChampionJulianna PeaRankingsRaquel PenningtonKayla HarrisonKetlen VieiraMacy ChiassonNorma DumontIrene AldanaMayra Bueno SilvaHolly HolmKarol RosaYana SantosMiesha TateJulia AvilaAilin PerezJacqueline CavalcantiChelsea ChandlerRight now, the most exciting fighters in the division are Pea, Harrison, Perez, Cavalcanti, and possibly Chandler. However, aside from the first two, no one else should represent a major impediment for a rising star.Well have to wait to see if Pereira can establish herself as a fighter who is ready to push toward a ranking sooner rather than later.

Today's NYT Strands hints and answersCredit: New York TimesLooking for Tuesdays Strands hints, spangram and answers? You can find them here:Its Christmas at last! I hope all you fine Strandsian puzzle-solvers have a lovely holiday and get to spend some quality time with your loved ones. Hopefully that quality time involves solving puzzles and playing games. I asked for a Christmas-themed Strands this year and Santa really delivered. Lets solve it!Strands is the newest game in the New York Times stable of puzzle games. Its a fun twist on classic word search games. Every day were given a new theme and then tasked with uncovering all the words on the grid that fit that theme, including a spangram that spans two sides of the board. One of these words is the spangram which crosses from one side of the grid to another and reveals even more about the days theme.Spoilers ahead.Todays Strands HintsTodays Theme: A visit from Santa.Hint #1: Think presents you stuff, not those your wrap.MORE FOR YOUHint #2: I dont think you need a second hint, frankly.To help you uncover all the words, here are the first two letters of every word, including the spangram.TOPLORSOSCCAPECOSTRemember, spoilers ahead!What Are Todays Strands Answers?Todays spangram is: STOCKINGHeres the full list of words:TOYSPLUSHIEORANGESOCKSSCARFCANDYPENSCOALHeres the completed Strands grid:Today's StrandsScreenshot: Erik KainTodays Strands BreakdownI found TOYS instantly and then noticed the ING next to it. Working backward from that, I pretty quickly found the spangram, and then it was just a matter of unscrambling the trickier words. CANDY and COAL and PENS were all simply enough. So were SOCKS and SCARF. The left side of the boardORANGE and PLUSHIE, in particularwere what stumped me the longest. Nine words total is a lot for a Strandsand a lot to stuff into a stocking. Merry Christmas, everyone!How did you do on your Strands today? Let me know on Twitter and Facebook.Be sure to check out my blog for my daily Wordle guides as well as all my other writing about TV shows, streaming guides, movie reviews, video game coverage and much more. Thanks for stopping by!

SpaceX has had its busiest year yet for rocket launches.The spaceflight company led by Elon Musk has so far conducted 129 Falcon 9 launches, two Falcon Heavy launches, and three Starship launches. Three more Falcon 9 missions are planned before the end of the year, making a grand total of 137 launches for 2024 way more than the 98 launches achieved in 2023.Recommended VideosMost of the rockets launched from the Kennedy Space Center in Florida, while some missions started from the Vandenberg Space Force Base in California. All three of the Starship flights launched from SpaceXs Starbase facility in Boca Chica, Texas.Please enable Javascript to view this contentThe Falcon rocket flights mostly involved deployments of satellites for Starlink, SpaceXs internet-from-space service. But there were also more than 20 launches for a range of commercial customers, up from 12 in 2023. Launches involving government contracts, meanwhile, ticked up to around 18 for the year.RelatedIt also conducted several rideshare missions, which keep costs down by serving multiple customers at once. One of these missions involved more than 100 payloads.It wasnt just satellites that SpaceX is deploying, as it also sent Crew-8 and Crew-9 to the space station, along with crew attached to the privately funded Axiom-3 mission. Additionally, SpaceX used a Falcon 9 to launch the privately funded Polaris Dawn mission, which sent a crew to the furthest point from Earth since the Apollo missions five decades, and also involved the first commercial spacewalk.As usual, SpaceX brought most of its first-stage Falcon 9 boosters home shortly after theyd deployed their payloads or astronauts to orbit. Only one Falcon 9 booster landing ended in failure this year. It happened in August when the vehicle caught fire during the touchdown, before toppling over.SpaceX has actually fallen slightly short of its original launch forecast for 2024 due to delays caused by a number of weather and technical issues.Looking ahead, 2025 is set to be even busier, with SpaceX aiming to serve more customers than ever. Its also aiming to ramp up testing of its mighty Starship rocket with as many as 25 launches taking place over the next 12 months. To date, the rocket has only flown six times, with the first one launching in April 2023. SpaceX already has permission for the Starships seventh test flight, which some reports suggest could take place in early January.Creating a record 17 million pounds of thrust as it leaves the ground, a Starship launch would mark a spectacular beginning to SpaceXs 2025 campaign.Editors Recommendations