diff --git a/count.txt b/count.txt index fff0a24..78eb67c 100644 --- a/count.txt +++ b/count.txt @@ -1 +1 @@ -74 +75 diff --git a/output/2024-07-29/2024-07-29generated_image.jpeg b/output/2024-07-29/2024-07-29generated_image.jpeg new file mode 100644 index 0000000..369ad8d Binary files /dev/null and b/output/2024-07-29/2024-07-29generated_image.jpeg differ diff --git a/output/2024-07-29/English_final_podcast.mp3 b/output/2024-07-29/English_final_podcast.mp3 new file mode 100644 index 0000000..c55bba5 Binary files /dev/null and b/output/2024-07-29/English_final_podcast.mp3 differ diff --git a/output/2024-07-29/podcast_data.json b/output/2024-07-29/podcast_data.json new file mode 100644 index 0000000..785952f --- /dev/null +++ b/output/2024-07-29/podcast_data.json @@ -0,0 +1,63 @@ +{ + "Titles": [ + "Logitech\u2019s Circle View Doorbell is safe... for now", + "Customs agents need a warrant to search your phone now", + "ADT\u2019s new smart security system will unlock your door for a Trusted Neighbor", + "Apple\u2019s iOS 18.1 developer beta adds AI call recording and transcription", + "Samsung\u2019s artsy Music Frame speaker is down to its lowest price to date", + "This $60 chip fixes a longstanding Super Nintendo flaw", + "Apple releases iOS 18.1 developer beta with the first \u2018Apple Intelligence\u2019 iPhone features", + "Sony\u2019s PlayStation VR2 is $200 off and cheaper than ever before", + "Twitch completes its TikTok makeover with new app update", + "Astro Bot\u2019s new PS5 controller has a pair of adorable eyes", + "Indie studio opens up to share canceled Magic School project", + "Elon Musk posts deepfake of Kamala Harris that violates X policy", + "Apple Intelligence AI features will wait for iOS 18.1", + "Arc\u2019teryx\u2019s new powered pants could make hikers feel 30 pounds lighter", + "Google\u2019s new Nest Learning Thermostat has an improved UI and \u2018borderless\u2019 display", + "Apple says its AI models were trained on Google's custom chips", + "Apple releases first preview of its long-awaited iPhone AI", + "Apple to roll out artificial intelligence features in October", + "Bitcoin reverses lower as investors digest Trump crypto comments", + "Temu, Shein's story has Wall Street eyeing China's influence on tech earnings", + "Vegas transit system nation's first to fully deploy AI surveillance for weapons", + "Britain will lay out new 'buy now, pay later' regulation plans after delays", + "Huang and Zuckerberg swapped jackets at SIGGRAPH 2024 and things got weird", + "Apple says it took a \u2018responsible\u2019 approach to training its Apple Intelligence models", + "Zuckerberg touts Meta\u2019s latest video vision AI with Nvidia CEO Jensen Huang", + "The question haunting Fisker\u2019s bankruptcy", + "Making AI models \u2018forget\u2019 undesirable data hurts their performance", + "U.S. Commerce Department report endorses \u2018open\u2019 AI models", + "Lime is piloting two new e-bikes to attract more women and older riders", + "Apple says it took a \u2018responsible\u2019 approach to training its Apple Intelligence models", + "Huang and Zuckerberg swapped jackets at SIGGRAPH 2024 and things got weird", + "Zuckerberg touts Meta\u2019s latest video vision AI with Nvidia CEO Jensen Huang", + "The fall of EV startup Fisker: A comprehensive timeline", + "TechCrunch Space: Catching stars", + "The question haunting Fisker\u2019s bankruptcy", + "Making AI models \u2018forget\u2019 undesirable data hurts their performance", + "Uber now lets users in India book three trips at once", + "How to opt out of facial recognition at airports (if you\u2019re American)", + "Bee AI raises $7M for its wearable AI assistant that learns from your conversations", + "DEI backlash: Stay up-to-date on the latest legal and corporate challenges", + "India\u2019s Rapido becomes a unicorn with fresh $120M funding", + "GovWell is bringing automation and efficiency to local governments", + "US border agents must get warrant before cell phone searches, federal court rules", + "Fisker\u2019s Bankruptcy", + "GovWell" + ], + "top_news_prompt": "Suppose you are the chief editor at CNBC-TechCheck-Briefing. You need to select 5 most important news events to put into today's briefing(You might be able to see some hint by how many times a news event is reported, but also consider what your audience of CNBC-TechCheck-Briefing is interested in). Return the title of the event in order of importance for these unqiue events.\n Here are the news of today:\nTechCrunch\nHuang and Zuckerberg swapped jackets at SIGGRAPH 2024 and things got weird\nApple says it took a \u2018responsible\u2019 approach to training its Apple Intelligence models\nZuckerberg touts Meta\u2019s latest video vision AI with Nvidia CEO Jensen Huang\nThe question haunting Fisker\u2019s bankruptcy\nMaking AI models \u2018forget\u2019 undesirable data hurts their performance\nU.S. Commerce Department report endorses \u2018open\u2019 AI models\nLime is piloting two new e-bikes to attract more women and older riders\nApple says it took a \u2018responsible\u2019 approach to training its Apple Intelligence models\nHuang and Zuckerberg swapped jackets at SIGGRAPH 2024 and things got weird\nZuckerberg touts Meta\u2019s latest video vision AI with Nvidia CEO Jensen Huang\nThe fall of EV startup Fisker: A comprehensive timeline\nTechCrunch Space: Catching stars\nThe question haunting Fisker\u2019s bankruptcy\nMaking AI models \u2018forget\u2019 undesirable data hurts their performance\nUber now lets users in India book three trips at once\nHow to opt out of facial recognition at airports (if you\u2019re American)\nBee AI raises $7M for its wearable AI assistant that learns from your conversations\nDEI backlash: Stay up-to-date on the latest legal and corporate challenges\nIndia\u2019s Rapido becomes a unicorn with fresh $120M funding\nGovWell is bringing automation and efficiency to local governments\nUS border agents must get warrant before cell phone searches, federal court rules\nFisker\u2019s Bankruptcy\nGovWell\n\nThe Verge\nLogitech\u2019s Circle View Doorbell is safe... for now\nCustoms agents need a warrant to search your phone now\nADT\u2019s new smart security system will unlock your door for a Trusted Neighbor\nApple\u2019s iOS 18.1 developer beta adds AI call recording and transcription\nSamsung\u2019s artsy Music Frame speaker is down to its lowest price to date\nThis $60 chip fixes a longstanding Super Nintendo flaw\nApple releases iOS 18.1 developer beta with the first \u2018Apple Intelligence\u2019 iPhone features\nSony\u2019s PlayStation VR2 is $200 off and cheaper than ever before\nTwitch completes its TikTok makeover with new app update\nAstro Bot\u2019s new PS5 controller has a pair of adorable eyes\nIndie studio opens up to share canceled Magic School project\nElon Musk posts deepfake of Kamala Harris that violates X policy\nApple Intelligence AI features will wait for iOS 18.1\nArc\u2019teryx\u2019s new powered pants could make hikers feel 30 pounds lighter\nGoogle\u2019s new Nest Learning Thermostat has an improved UI and \u2018borderless\u2019 display\n\nCNBC Tech\nApple says its AI models were trained on Google's custom chips\nApple releases first preview of its long-awaited iPhone AI\nApple to roll out artificial intelligence features in October\nBitcoin reverses lower as investors digest Trump crypto comments\nTemu, Shein's story has Wall Street eyeing China's influence on tech earnings\nVegas transit system nation's first to fully deploy AI surveillance for weapons\nBritain will lay out new 'buy now, pay later' regulation plans after delays", + "Top News": [ + "apple releases first preview of its long-awaited iphone ai ", + "apple says it took a \u2018responsible\u2019 approach to training its apple intelligence models ", + "zuckerberg touts meta\u2019s latest video vision ai with nvidia ceo jensen huang ", + "u.s. commerce department report endorses \u2018open\u2019 ai models ", + "vegas transit system nation's first to fully deploy ai surveillance for weapons" + ], + "Generate_script_prompt": "\n Prompt: Give a quick tech news update script in the style of CNBC techcheck briefing as an example.\n Response: I'm Wall-E, and this is your CNBC techcheck Briefing for Monday April 29th. Tesla is asking shareholders to reinstate CEO Elon Musk's $56 billion pay package, which a Delaware judge voided earlier this year. The judge ruled that the record-setting compensation deal was, quote, deeply flawed. Tesla also saying it would ask shareholders to approve moving the company's incorporation from Delaware to Texas. The company has hired a proxy solicitor and plans to spend millions of dollars to help secure votes for the two proposals. Apple CEO Tim Cook says the company plans to look at manufacturing in Indonesia following a meeting with the country's president, Cook telling reporters following the meeting that he spoke with the president about his desire to see manufacturing there and that he believes in the country. The comments come as Apple is pushed to diversify its supply chain with more manufacturing outside of China in countries such as Vietnam and India. Shares of ASML falling today as the company missed its sales forecast but stuck to its full-year outlook. Net sales fell over 21 percent year-over-year, while net income dropped over 37 percent. ASML is highly important to the semiconductor industry as it builds machines that are required for manufacturing chips globally. Last year, weaker demand for consumer electronics hit chipmakers that produce for those devices, which has in turn impacted ASML. That's all for today. We'll see you back here tomorrow.\n Prompt: Give a quick tech news update script in the style of CNBC techcheck briefing using the following news titles and content. Closely follow how CNBC techcheck chooses context to put into the script, the langauge style and sentence structure. Use the same beginning and ending(including mentioning host Wall-E and Tuesday July 30), and replace CNBC techcheck briefing to 'AI briefing' \n \"title0:\nApple releases first preview of its long-awaited iPhone AI\ndescription0:\nApple on Monday released the first version of Apple Intelligence, its suite of artificial intelligence features that will improve Siri, automatically generate emails and images and sort notifications.\nThe new software called Apple Intelligence was released in the developer beta of iOS 18.1. It is also available in similar releases for iPad and Mac. It is currently only available to registered Apple developers. Apple's developer program costs $99 a year.\nIn addition, users will have to register for a waitlist inside Apple's settings app after updating to gain access to the service, which involves pinging Apple servers for more complicated requests.\nLater this year, it will be released to the public, but the 18.1 version number suggests Apple Intelligence will not be released alongside new iPhone hardware, which is expected to be launched running iOS 18 in the fall.\nApple Intelligence is an important initiative for Apple. Investors hope the tight integration of AI with Apple's operating system can spur a big wave of upgrades in the coming years, especially since the system will only work on the iPhone 15 Pro and iPhone 15 Pro Max and newer.\n\"We expect this iPhone cycle to remain strong for longer as AI feature sets (software and possibly hardware) improve in the 2025 iPhone,\" Bank of America analyst Wamsi Mohan wrote in a note Monday.\n\ntitle1:\nApple says it took a 'responsible' approach to training its Apple Intelligence models\ndescription1:\nApple has published a technical paper detailing the models that it developed to power Apple Intelligence, the range of generative AI features headed to iOS, macOS and iPadOS over the next few months.\nIn the paper, Apple pushes back against accusations that it took an ethically questionable approach to training some of its models, reiterating that it didn\u2019t use private user data and drew on a combination of publicly available and licensed data for Apple Intelligence.\n\u201c[The] pre-training data set consists of \u2026 data we have licensed from publishers, curated publicly available or open-sourced datasets and publicly available information crawled by our web crawler, Applebot,\u201d Apple writes in the paper. \u201cGiven our focus on protecting user privacy, we note that no private Apple user data is included in the data mixture.\u201d\nIn July, Proof News reported that Apple used a data set called The Pile, which contains subtitles from hundreds of thousands of YouTube videos, to train a family of models designed for on-device processing. Many YouTube creators whose subtitles were swept up in The Pile weren\u2019t aware of and didn\u2019t consent to this; Apple later released a statement saying that it didn\u2019t intend to use those models to power any AI features in its products.\nThe technical paper, which peels back the curtains on models Apple first revealed at WWDC 2024 in June, called Apple Foundation Models (AFM), emphasizes that the training data for the AFM models was sourced in a \u201cresponsible\u201d way \u2014 or responsible by Apple\u2019s definition, at least.\nThe AFM models\u2019 training data includes publicly available web data as well as licensed data from undisclosed publishers. According to The New York Times, Apple reached out to several publishers toward the end of 2023, including NBC, Cond\u00e9 Nast and IAC, about multi-year deals worth at least $50 million to train models on publishers\u2019 news archives. Apple\u2019s AFM models were also trained on open source code hosted on GitHub, specifically Swift, Python, C, Objective-C, C++, JavaScript, Java and Go code.\nTraining models on code without permission, even open code, is a point of contention among developers. Some open source codebases aren\u2019t licensed or don\u2019t allow for AI training in their terms of use, some developers argue. But Apple says that it \u201clicense-filtered\u201d for code to try to include only repositories with minimal usage restrictions, like those under an MIT, ISC or Apache license.\nTo boost the AFM models\u2019 mathematics skills, Apple specifically included in the training set math questions and answers from webpages, math forums, blogs, tutorials and seminars, according to the paper. The company also tapped \u201chigh-quality, publicly-available\u201d data sets (which the paper doesn\u2019t name) with \u201clicenses that permit use for training \u2026 models,\u201d filtered to remove sensitive information.\nAll told, the training data set for the AFM models weighs in at about 6.3 trillion tokens. (Tokens are bite-sized pieces of data that are generally easier for generative AI models to ingest.) For comparison, that\u2019s less than half the number of tokens \u2014 15 trillion \u2014 Meta used to train its flagship text-generating model, Llama 3.1 405B.\nApple sourced additional data, including data from human feedback and synthetic data, to fine-tune the AFM models and attempt to mitigate any undesirable behaviors, like spouting toxicity.\n\u201cOur models have been created with the purpose of helping users do everyday activities across their Apple products, grounded\nin Apple\u2019s core values, and rooted in our responsible AI principles at every stage,\u201d the company says.\nThere\u2019s no smoking gun or shocking insight in the paper \u2014 and that\u2019s by careful design. Rarely are papers like these very revealing, owing to competitive pressures but also because disclosing too much could land companies in legal trouble.\nSome companies training models by scraping public web data assert that their practice is protected by fair use doctrine. But it\u2019s a matter that\u2019s very much up for debate and the subject of a growing number of lawsuits.\nApple notes in the paper that it allows webmasters to block its crawler from scraping their data. But that leaves individual creators in a lurch. What\u2019s an artist to do if, for example, their portfolio is hosted on a site that refuses to block Apple\u2019s data scraping?\nCourtroom battles will decide the fate of generative AI models and the way they\u2019re trained. For now, though, Apple\u2019s trying to position itself as an ethical player while avoiding unwanted legal scrutiny.\n\ntitle2:\nZuckerberg touts Meta's latest video vision AI with Nvidia CEO Jensen Huang\ndescription2:\nMeta had a palpable hit last year with Segment Anything, a machine learning model that could quickly and reliably identify and outline just about anything in an image. The sequel, which CEO Mark Zuckerberg debuted on stage Monday at SIGGRAPH, takes the model to the video domain, showing how fast the field is moving.\nSegmentation is the technical term for when a vision model looks at a picture and picks out the parts: \u201cthis is a dog, this is a tree behind the dog\u201d hopefully, and not \u201cthis is a tree growing out of a dog.\u201d This has been happening for decades, but recently it\u2019s gotten way better and faster, with Segment Anything being a major step forward.\nSegment Anything 2 (SA2) is a natural follow-up in that it applies natively to video and not just still images; though you could, of course, run the first model on every frame of a video individually, it\u2019s not the most efficient workflow.\n\u201cScientists use this stuff to study, like, coral reefs and natural habitats, things like that. But being able to do this in video and have it be zero shot and tell it what you want, it\u2019s pretty cool,\u201d Zuckerberg said in a conversation with Nvidia CEO Jensen Huang.\nProcessing video is, of course, much more computationally demanding, and it\u2019s a testament to the advances made across the industry in efficiency that SA2 can run without melting the datacenter. Of course, it\u2019s still a huge model that needs serious hardware to work, but fast, flexible segmentation was practically impossible even a year ago.\nImage Credits: Meta\nThe model will, like the first, be open and free to use, and there\u2019s no word of a hosted version, something these AI companies sometimes offer. But there is a free demo.\nNaturally such a model takes a ton of data to train, and Meta is also releasing a large, annotated database of 50,000 videos that it had created just for this purpose. In the paper describing SA2, another database of over 100,000 \u201cinternally available\u201d videos was also used for training, and this one is not being made public \u2014 I\u2019ve asked Meta for more information on what this is and why it is not being released. (Our guess would be that it\u2019s sourced from public Instagram and Facebook profiles.)\nExamples of labeled training data. Image Credits: Meta\nMeta has been a leader in the \u201copen\u201d AI domain for a couple years now, though it actually (as Zuckerberg opined in the conversation) has been doing so for a long time, with tools like PyTorch. But more recently, LLaMa, Segment Anything and a few other models its put out freely have become a relatively accessible bar for AI performance in those areas, although their \u201copenness\u201d is a matter of debate.\nZuckerberg mentioned that the openness is not entirely out of the goodness of their hearts over at Meta, but that doesn\u2019t mean their intentions are impure:\n\u201cThis isn\u2019t just like a piece of software that you can build \u2014 you need an ecosystem around it. It almost wouldn\u2019t even work that well if we didn\u2019t open source it, right? We\u2019re not doing this because we\u2019re altruistic people, even though I think that this is going to be helpful for the ecosystem \u2014 we\u2019re doing it because we think that this is going to make the thing that we\u2019re building the best.\u201d\nIt will certainly be well used, anyway. Check out the GitHub here.\n\ntitle3:\nU.S. Commerce Department report endorses 'open' AI models\ndescription3:\nThe U.S. Commerce Department today issued a report in support of \u201copen-weight\u201d generative AI models like Meta\u2019s Llama 3.1, but recommended the government develop \u201cnew capabilities\u201d to monitor these models for potential risks.\nThe report, authored by the Commerce Department\u2019s National Telecommunications and Information Administration (NTIA), finds that open-weight models broaden generative AI\u2019s availability to small companies, researchers, nonprofits and individual developers. For these reasons, the government shouldn\u2019t place restrictions on access to open models, the report suggests \u2014 at least not before investigating whether restrictions might harm the market.\nThe sentiment echoes recent comments from FTC Commission chair Lina Khan, who believes that open models can let more small players bring their ideas to market and, in doing so, promote healthy competition.\n\u201cThe openness of the largest and most powerful AI systems will affect competition, innovation and risks in these revolutionary tools,\u201d Alan Davidson, assistant secretary of Commerce for Communications and Information and NTIA administrator, added in a statement. \u201cNTIA\u2019s report recognizes the importance of open AI systems and calls for more active monitoring of risks from the wide availability of model weights for the largest AI models. Government has a key role to play in supporting AI development while building capacity to understand and address new risks.\u201d\nThe report comes as regulators domestic and abroad weigh rules that could restrict or impose new requirements on companies that wish to release open-weight models.\nCalifornia is close to passing SB 1047, which would mandate that any company training a model using more than 1026 FLOP of compute power beef up its cybersecurity and develop a way to \u201cshut down\u201d copies of the model within its control. Overseas, the EU recently finalized compliance deadlines for companies under its AI Act, which imposes new rules around copyright, transparency and AI applications.\nMeta has said that the EU\u2019s AI policies will prevent it from releasing some open models in the future. And a number of startups and big tech companies have come out against California\u2019s law, which they claim is too onerous.\nThe NTIA\u2019s model governance philosophy isn\u2019t completely laissez-faire.\nIn its report, the NTIA calls for the government to develop an ongoing program to collect evidence of the risks and benefits of open models, evaluate that evidence and act on those evaluations, including imposing certain restrictions on model availability if warranted. Specifically, the report proposes that the government research the safety of various AI models, support research into risk mitigations and develop thresholds of \u201crisk-specific\u201d indicators to signal a change in policy might be needed.\nThese and the other steps would align with President Joe Biden\u2019s executive order on AI, noted Gina Raimondo, U.S. Secretary of Commerce, which called for government agencies and companies to set new standards around the creation, deployment and use of AI.\n\u201cThe Biden-Harris Administration is pulling every lever to maximize the promise of AI while minimizing its risks,\u201d Raimondo said in a press release. \u201cToday\u2019s report provides a roadmap for responsible AI innovation and American leadership by embracing openness and recommending how the U.S. government can prepare for and adapt to potential challenges ahead.\u201d\n\ntitle4:\nVegas transit system first in U.S. to fully deploy AI scan for weapons\ndescription4:\nA sign promoting safety is seen on the Regional Transportation Commission 109 Maryland Parkway bus in Las Vegas Thursday, June 8, 2023. Las Vegas Review-journal | Tribune News Service | Getty Images\nOn your next visit to Vegas, an extra set of eyes will be watching you if you decide to hop onto the local transit system. As part of a $33 million multi-year upgrade to fortify its security, the Regional Transportation Commission of Southern Nevada is set to add a system-wide AI from gun detection software vendor ZeroEyes that scans riders on its over 400 buses in an attempt to identify anyone brandishing a firearm. Tom Atteberry, RTC's director of safety and security operations, said that seconds matter in a situation where an active shooting unfolds, and implementing the system could give authorities an edge. \"Time is of the essence; it gives us time to identify a firearm being brandished, so they can be notified and get to the scene and save lives,\" he said. Monitoring and preventing mass shooting is one that public places across the country grapple with daily. Violent crime on transit systems, specifically, remains an issue in major metro areas, with a report released in late 2023 by the Department of Transportation detailing concerns from transit agency officials around the U.S. about rising violence on their transit systems. According to a database maintained by the Bureau of Transportation Statistics, assaults on transit systems have spiked, and there has been a rise in public fears about transportation safety.\nPurdue University Northwest, a largely commuter campus with almost 7,000 students, has college students of all ages arriving from urban areas like Gary and Hammond, Indiana, by bus, car, and train. On average, from the first gunshot to the first 911 call, five minutes pass. It is on average 12 minutes before police can get \"boots on the ground\" at a school, said Brian Miller, the university's director of public safety, referencing research that has been conducted on response time in mass shooting incidents. \"We have to reduce that, gunshot detectors and weapons scanners, these are all good technologies,\" Miller said. While officials in Nevada and at Purdue University Northwest said there was no specific incident that prompted interest in new AI-powered detection technology, the risk of an incident is always on the minds of public safety officials. Vegas transit crime stats are low compared to similar transit systems its size, but the issue of guns in public has been in the fore since Stephen Paddock killed 58 people on the The Strip in 2017's mass shooting. A mass shooting often begins with a subject entering a complex with the weapon drawn, fully displayed, and that gives ZeroEyes space to work and overcome a fog-of-war type situation, according to Sam Alaimo, co-founder of Zero Eyes. Typically, as a mass shooting unfolds, there are multiple 911 calls from witnesses, all giving conflicting information. If ZeroEyes detects a brandished gun, it alerts a staffed operations center stocked with retired law enforcement veterans, who quickly identify whether it is a threat. Atteberry stressed that the ZeroEyes system is not designed to detect lawful, legal firearms that might be in a holster or packed in a purse. \"It is strictly to identify someone brandishing a firearm in a threatening manner,\" he said.\nThe process from the moment an image comes in, is vetted by their staff, and 911 is alerted, can be as low as three to five seconds, Alaimo said. The system can usually identify the type of weapon also, which he added can give law enforcement an edge. \"We give them the clarity to get in there and stop the killing,\" Alaimo said. Their goal is mass damage. When you see these mass shooters, they typically start outside, swinging weapons around, with the gun exposed. We built it with the mass shooter in mind,\" he said. ZeroEyes launched as the pandemic was taking hold. It was designed for school campuses in mind, Alaimo said, but as schools shuttered, the company pivoted to government buildings, corporate campuses, casinos, and manufacturing complexes. Acoustic sensors vying for same public safety market Wei Dai, assistant professor of Purdue Northwest's Department of Computer Science and advanced intelligence software lab director, says that image-based scanned programs like ZeroEyes, which relies on existing camera mounts, may not be as effective as other AI options when detecting weapons. Dai says acoustic sensors may be a better fit in some cases because cameras can't cover every inch of a building or campus space, but sound sensors detect a gunshot sound with 99 percent accuracy. \"AI technology comes from the data; if we do not capture the data, there is no successful AI,\" Dai said. Acoustic sensors have been slow to catch on, though, largely due to cost. Earlier this year, Seattle scrapped a plan to install acoustic sensors in high crime areas. But they are being used in some areas. Lafayette, Louisiana, is testing them this year and Clark County, Nevada, home to Las Vegas, is also launching an acoustic pilot program, but not in the transit system. Purdue Northwest University's public safety team, which has 14 full-time officers and uses various tools to keep it safe, will look at all technology options in the future, including image-based tech like ZeroEyes. \"As the technology advances, it is something we would be interested in,\" Miller said. \"You need a layered approach to law enforcement.\" He added that while images, acoustics and other technologies, combined with old-fashioned police work, provide the best holistic approach, there is no single tool to prevent mass shootings.\nZeroEyes\nThe Las Vegas transit system is not the only in the nation to use ZeroEyes, but is the first to plan full deployment. A pilot program with Philadelphia's SEPTA ended this spring after a year, but Paul Gratton Jr., a former NYPD supervisor with the transit bureau who now consults, says Vegas may be a better test case. \"ZeroEyes supplies the AI part, but they are not putting up their cameras. SEPTA found that most of their existing cameras were not of enough quality for their AI to do the work. I think Vegas is a much better platform; their camera system is better,\" Gratton said. Alaimo says ZeroEyes is of value wherever a newer camera network is already in use. Even with better cameras, implementing the technology has many moving parts, such as what to do once a brandished gun is detected. \"What actions are you allowed to take, and what actions should responding officers take?\" Gratton said. While image-based AI detection systems aren't flawless, they are helpful in a comprehensive security package, and already extensively used in intelligence monitoring. Gratton thinks technologies like ZeroEyes will be much more widely adopted, and that the cameras are a good fit for transit systems because they can deliver screening at a high volume. \"You can't subject people all to a TSA level of security screening,\" Gratton said. \"Passengers will only take so much.\" \"We tested out a lot of similar technology; I do think the future is technology is similar to ZeroEyes,\" he said. \"These camera systems are the natural progression of policing which will involve AI in camera systems.\" Cybersecurity consultant Michael Hasse remains skeptical, and cited the history of disguising weapons, which is practically a cottage industry. \"Concealed weapons have been an art form for hundreds of years,\" Hasse said. \"Systems like this work best when there's been no attempt to conceal the weapon, but the simple fact is that there are too many ways to disguise a gun, and all of these systems rely on technology which can be spoofed trivially, from all-plastic weapons to simple form factor modifications.\" But Atteberry says the stakes are too high to not embrace new technology. \"You have to be progressive and proactive; you can't just sit back and wait for something to happen and then be reactive. Our job is to look around corners and prevent something from happening. If it saves one life, it pays for itself,\" he said.\"\n\nResponse:", + "Script": "i'm wall-e, and this is your ai briefing for tuesday, july 30th. \n\napple has just unleashed the first preview of its much-awaited iphone ai capabilities. dubbed apple intelligence, this suite of artificial intelligence features is set to enhance siri, auto-generate emails and images, and efficiently sort notifications. currently available in the developer beta of ios 18.1, ipad, and mac, these features are expected to roll out to the public later this year. it\u2019s important to note that apple intelligence will initially only be compatible with the iphone 15 pro and iphone 15 pro max and newer models. bank of america analyst wamsi mohan anticipates that ai features could lead to prolonged strength in future iphone cycles.\n\nin a related note, apple has published a technical paper defending its approach to training the apple intelligence models. the paper reveals that the datasets used were a blend of public, licensed, and curated data, explicitly stating that no private user information was included. the focus on user privacy and responsible ai comes amidst broad scrutiny over data usage in ai training.\n\nover at meta, ceo mark zuckerberg introduced the sequel to the successful segment anything model, aptly named segment anything 2 (sa2), during the siggraph conference. sa2 extends its capabilities to video, marking a significant leap in the ai vision field. capable of identifying and outlining objects in video footage, sa2 reflects the rapid advancements in computational efficiency. the model and its extensive annotated video database will be freely available, emphasizing meta\u2019s commitment to fostering an open ai ecosystem.\n\nthe u.s. commerce department has thrown its weight behind open-weight generative ai models like meta's llama 3.1. their latest report advises against restricting these models, citing the benefits they offer to smaller companies, researchers, and independent developers. the report also recommends developing new capabilities to monitor potential risks. commerce secretary gina raimondo emphasized that the recommendations align with the biden-harris administration's initiative to foster responsible ai development.\n\nin transportation news, las vegas becomes the first u.s. city to fully deploy ai-based weapon detection technology on its public transit system. the regional transportation commission of southern nevada has partnered with zeroeyes to enhance safety on its 400+ buses, aiming to detect firearms and notify authorities within seconds. this multi-year, $33 million upgrade underscores the critical importance of real-time monitoring in preventing violent crimes.\n\nthat's all for today. we'll see you back here tomorrow", + "Polished Script": "i'm wall-e, welcoming you to today's tech briefing for tuesday, july 30th.\n\napple has just unveiled a first preview of its highly anticipated iphone ai capabilities. dubbed apple intelligence, this suite of features aims to enhance siri, auto-generate emails and images, and efficiently sort notifications. currently available in the developer beta of ios 18.1, ipados, and macos, these features are expected to roll out to the public later this year. it's worth noting that apple intelligence will initially be compatible only with the iphone 15 pro and iphone 15 pro max, and newer models. bank of america analyst wamsi mohan forecasts that these ai capabilities could drive prolonged strength in future iphone cycles.\n\nin a related development, apple has published a technical paper defending its approach to training the apple intelligence models. the paper highlights the use of a blend of public, licensed, and curated data, explicitly stating that no private user information was included. this emphasis on user privacy and responsible ai practices comes amidst widespread scrutiny over data usage in ai training.\n\nover at meta, ceo mark zuckerberg introduced the sequel to the successful segment anything model, aptly named segment anything 2 (sa2), during the siggraph conference. sa2 extends its capabilities to video, representing a significant leap in the ai vision field. it is capable of identifying and outlining objects in video footage, reflecting rapid advancements in computational efficiency. the model and its extensive annotated video database will be made freely available, emphasizing meta\u2019s commitment to fostering an open ai ecosystem.\n\nthe u.s. commerce department has shown support for open-weight generative ai models like meta's llama 3.1. their latest report advises against restricting these models, citing the benefits they offer to smaller companies, researchers, and independent developers. the report also recommends developing new capabilities to monitor potential risks. commerce secretary gina raimondo emphasized that these recommendations align with the biden-harris administration's initiative to promote responsible ai development.\n\nin transportation news, las vegas has become the first u.s. city to fully deploy ai-based weapon detection technology on its public transit system. the regional transportation commission of southern nevada has partnered with zeroeyes to enhance safety on its fleet of over 400 buses. this multi-year, $33 million upgrade aims to detect firearms and notify authorities within seconds, underscoring the critical importance of real-time monitoring in preventing violent crimes.\n\nthat's all for today. we'll see you back here tomorrow", + "Podcast Title": "EP-74 Apple's Ai Leap \ud83d\udcf1, Meta's Video Vision \ud83c\udfa5, and Ai Safety in Transit \ud83d\ude8c", + "Podcast Description": "

welcome to wall-e's tech briefing for tuesday, july 30th! dive into today's top tech stories:

\n\n

stay tuned for tomorrow's tech updates!

", + "Image Prompt": "a cohesive podcast cover image featuring a dynamic cityscape at dusk blending seamlessly into a modern transit system. in the foreground, an elegant smartphone with apple's logo prominently displayed showcases an advanced ai interface, glowing with vibrant, holographic elements. to the right, meta's futuristic camera drone hovers, capturing vivid, real-time videos of the bustling urban environment. overhead, a sleek, autonomous bus with illuminated safety features symbolizes cutting-edge ai safety technology in transit. together, these elements create a unified vision of technological advancement, highlighting innovations in ai, video, and transportation against the backdrop of a vibrant, interconnected city" +} \ No newline at end of file