Not only is AI-driven natural language processing a thing now, but you can even select from a number of different offerings, each optimized for different tasks. It took very little time for [Bloc] to mod a computer game to allow the player to converse naturally with non-player characters (NPCs) by hooking it into ChatGPT, a large language model AI optimized for conversational communication.
If you can look past the painfully-long loading times, even buying grain (7:36) gains a new layer of interactivity.
[Bloc] modified the game Mount & Blade II: Bannerlord to reject traditional dialogue trees and instead accept free-form text inputs, using ChatGPT on the back end to create more natural dialogue interactions with NPCs. This is a refinement of an earlier mod [Bloc] made and shared, so what you see in the video below is quite a bit more than a proof of concept. The NPCs communicate as though they are aware of surrounding events and conditions in the game world, are generally less forthcoming when talking to strangers, and the new system can interact with game mechanics and elements such as money, quests, and hirelings.
Starting around 1:08 into the video, [Bloc] talks to a peasant about some bandits harassing the community, and from there demonstrates hiring some locals and haggling over prices before heading out to deal with the bandits.
The downside is that ChatGPT is currently amazingly popular. As a result, [Bloc]’s mod is stuck using an overloaded service which means some painfully-long load times between each exchange. But if you can look past that, it’s a pretty fascinating demonstration of what’s possible by gluing two systems together with a mod and some clever coding.
Google’s search engine is getting an upgrade—yes with AI, but also with a blur filter. In a bid to shield its user’s eyes, Google announced yesterday that it will blur explicit images in search results, even when SafeSearch isn’t turned on.
The news comes in a blog post published by Google Senior Vice President of Core Systems & Experiences, Jen Fitzpatrick. While SafeSearch already filters out unsavory results for users that are logged in and under the age of 18, the blur setting will apply to explicit search results even when SafeSearch is turned off. According to the post, the feature will give searchers the option to view the image, but with the warning that it may contain explicit content. Fitzpatrick says that users can opt out of this feature in their SafeSearch settings.
[…]
This is actually not Google’s first time flirting with blurring potentially sensitive content. In December, the company announced that Magritte, a similar blurring tool, was going open source. Magritte won’t blur entire images necessarily, but can blur certain parts of images, like faces and license plates.
Microsoft’s oft-forgotten search engine is about to get a new lease on life. As expected, the tech giant announced on Tuesday that it is partnering with OpenAI to enhance Bing with the company’s AI technology. However, Microsoft also had a surprise up its sleeve: the next release of Bing will feature a new next-generation large language model the company claims is “much more powerful” than ChatGPT and designed specifically to excel at search.
Underpinning the search engine is a new proprietary technology Microsoft is calling the Prometheus Model. Among the benefits of Prometheus are more relevant search results. Microsoft claims the model will also make using Bing safer and allow the company to update search results more quickly.
Microsoft
The new Bing offers a chat function and an expanded search bar that allows you to input up to a thousand characters. In one demo Microsoft showed off during its event, Yusuf Mehdi, the company’s consumer chief marketing officer, asked Bing to compare the most influential Mexican artists and their best-known paintings. Bing displayed its response in a new side panel that featured annotations and weblinks. Later, Mehdi asked the search engine to compare three pet vacuums while listing the pros and cons of each model. Mehdi said Bing could also help with questions where there might not be a precise answer. That said, in instances where Bing doesn’t know the answer to a query, the search engine will still try to point you to useful information. Microsoft is also adding a button to allow users to share feedback on the search engine’s new functionality. “I think of this as Search with your own personal helper,” Mehdi said.
The chat feature is there for when you want to use the generative capabilities of OpenAI’s technology. For example, say you want to start planning a trip to Hawaii. You can ask Bing to create a five-day itinerary for you. In its response, the search engine will include links to accommodations, flights and things you can do on the islands. The new Bing is available to preview starting today. Microsoft says anyone can visit Bing.com to try a few sample queries and sign up for the company’s waitlist. “We’re going to scale the preview to millions in the coming weeks,” Microsoft said. A mobile preview is coming soon as well.
Now that Microsoft has launched its Teams Premium service, it’s shaking up the free offering for work — and not everyone will be happy. The company is retiring the existing Teams Free version for small business in favor of the similarly-titled Teams (free) on April 12th, and legacy data won’t carry over. Your office will have to pay for at least the Teams Essentials plan ($4 per user per month) to preserve chats, meetings, channels and other key info.
As Windows Centralexplains, the new Teams (free) tier will require a new account. Data in the old app, now rebadged as Teams Free (classic), will be deleted. Anything you haven’t saved by then will be gone, including shared files you haven’t downloaded.
We’ve asked Microsoft for comment. This won’t affect personal use, but it could prove to be a headache for small firms that previously relied on the free Teams to coordinate. They’ll either have to start paying or they’ll lose access to past discussions, not to mention deal with the headache of recreating their channel setups.
Microsoft’s use of AI on the web will extend beyond smarter Bing searches. The company is updating its Edge browser with an “AI-powered copilot” that assists with everyday tasks. Integration with the new Bing can help you get the key takeaways from a document, compare content and even draft posts (such as on LinkedIn) with a specific format, length or tone.
The revised interface revolves around a new sidebar. You can use a chat feature in this space to ask questions like you would with Bing, while a compose option helps you write new content using prompts. You can also expect a new look.
Microsoft
Both Bing and the updated Edge are based around a new AI model, Prometheus, built with the help of OpenAI. While it draws some lessons from the popular ChatGPT conversational AI, it’s billed as more powerful and optimized for search. Microsoft also claims it’s a “responsible” platform where protections against abuse and errors exist at every level.
The new Edge is available now in preview form. While Bing is the highlight of Microsoft’s AI announcements, the new browser clearly plays an important role. The AI capabilities theoretically give Edge an advantage over Chrome, Safari and other browsers. Of course, how long any lead lasts is uncertain. Google is holding an event tomorrow (February 8th) that may shed light on its bid to rival ChatGPT, and it won’t be surprising if AI-driven search is part of that strategy.
[…] According to outage tracker DownDetector, reports began coming in of users facing a 500 error and being unable to send, receive or search email through Outlook.com from about 4am UTC, peaking at 8 and 9am as Europeans reached their desks.
Microsoft confirmed the outage on its service health website, saying: “We’re applying targeted mitigations to a subset of affected infrastructure and validating that it has mitigated impact. We’re also making traffic optimization efforts to alleviate user impact and expedite recovery.”
It added that extra “Outlook.com functionality such as Calendar APIs consumed by other services such as Microsoft Teams are also affected.”
At the time of writing, the blackout appears to be ongoing. As for what caused it, the Microsoft 365 Status Twitter account said: “We’ve confirmed that a recent change is contributing to the cause of impact. We’re working on potential solutions to restore availability of the service.”
In plain English, Microsoft tweaked something and the house of cards came tumbling down, so they’ll probably have to revert the change. It offered the reference number EX512238 to track in the admin center and otherwise directed users to watch the service health page for any updates.
The EU’s upcoming artificial intelligence legislation will define how the world regulates AI, warned GitHub CEO Thomas Dohmke, who thinks open source developers should be exempted from the AI Act.
Speaking at the Open Source Policy Summit in Brussels (at 04:05:00), the day before FOSDEM in the Belgian city, Dohmke noted that open source powers governments and corporations alike and is at the “heart of European values” because of its belief that “everything can be better with collaboration.”
Making the case for lawmakers to take it easy on open sourcers, Dohmke said it was powering a “spring of innovation across the world and here in Europe,” seeing it as GitHub’s responsibility as the largest source code host to protect devs’ rights to “innovation and equal opportunity.”
The CEO said the policy could “well set a precedent for how the world regulates AI. It must be fair and balanced to the open source community.” He added:
The compliance burden should fall on companies that are shipping products. OSS developers are often just volunteers. Many of them are working two jobs, they are hobbyists and scientists, academics and doctors, professors and university students alike, and they don’t usually stand to profit from their contributions. They certainly don’t have big budgets or their own compliance department.
Dohmke was speaking as EU lawmakers reportedly look towards achieving a consensus on draft AI rules – the first of their kind globally – as soon as early March and against the backdrop of increasing regulation against Big Tech in Europe.
Dragos Tudorache, co-rapporteur of the EU AI Act (one of the lawmakers leading the committee and coordinating decision-making), reportedly told Reuters: “We are still in good time to fulfill the overall target and calendar that we assumed in the very beginning, which is to wrap it up during this mandate,” meaning the end of this year.
It will come as no surprise to developers working in the field that “this text has seen a level of complexity that is even higher than the typical Brussels complex machinery.”
The law was first proposed in 2021 and is something of a legislative jewel for EU parliamentarians because the issue hasn’t been comprehensively addressed in law before. There’s also the small matter of nationalism being a thing again, despite the world’s best efforts. Competition with the US and China – both home to companies frenziedly developing the tech – is fierce.
Those looking to influence the draft laws are in two camps: one that says you can’t do enough to address “risks” from AI systems, and the other believing that stricter laws can stifle creativity and innovation. The latter is home to both corporate devs and open source types.
There is a tension here, with American think tank Brookings pointing out last year that the new law could prevent developers from releasing open source models, considering their own liability, thus ensuring the development of AI would be driven by large corporations
