As fires rage across Los Angeles and tens of thousands flee their homes, the usual suspects have decided to blame the blazes on their political enemies. In a series of posts on Truth Social, President-elect Donald Trump claimed firefighters’ inability to get the fires under control was due to California Governor Gavin Newsom’s water policies, including an effort to “protect an essentially worthless fish called a smelt, by giving it less water (it didn’t work!).” Meanwhile, on X, Elon Musk suggested that the fires were spreading due to the city fire chief’s commitment to diversity, equity, and inclusion policies. “DEI means people DIE,” Musk wrote in a Wednesday night post.

Five people have indeed died in the Eaton fire thus far, and upward of 130,000 Los Angeles County residents are under evacuation orders, according to the Los Angeles Times. But the rampant spread of the fires isn’t due to the delta smelt, DEI, or even — as Trump, Musk, and scores of mainstream publications have falsely claimed — cuts to the Los Angeles Fire Department’s budget. These claims aren’t without consequence. Last year, FEMA workers received threats on TikTok and other social media platforms as rampant disinformation spread in the wake of the devastation caused by hurricanes Milton and Helene.

The first wave of disinformation focused on fire hydrants in the Pacific Palisades, which abruptly ran out of water on Tuesday night as firefighters attempted to put out the initial blaze. The hydrants, Trump said, were running dry because of Newsom’s water policies. “I will demand that this incompetent governor allow beautiful, clean, fresh water to FLOW INTO CALIFORNIA!” Trump posted on Truth Social on Wednesday. “He is the blame for this. On top of it all, no water for fire hydrants, not firefighting planes. A true disaster!” In a separate post, Trump claimed Newsom had “refused to sign the water restoration declaration put before him that would have allowed millions of gallons of water, from excess rain and snow melt from the North, to flow daily into many parts of California, including the areas that are currently burning in a virtually apocalyptic way.”

Newsom’s office responded on Wednesday, clarifying that the declaration Trump referred to in his post didn’t exist. “There is no such document as the water restoration declaration – that is pure fiction,” Newsom communications director Izzy Gardon told CalMatters. “The Governor is focused on protecting people, not playing politics, and making sure firefighters have all the resources they need.”

Mark Gold, a board member of the Metropolitan Water District of Southern California, told the outlet that statewide water management policies, including efforts to protect the delta smelt, had nothing to do with the lack of water in the hydrants. “Tying Bay-Delta management into devastating wildfires that have cost people’s lives and homes is nothing short of irresponsible, and it’s happening at a time when the Metropolitan Water District has the most water stored in its system in the history of the agency,” Gold said. “It’s not a matter of having enough water coming from Northern California to put out a fire. It’s about the continued devastating impacts of a changing climate.”

The lack of water in the Pacific Palisades hydrants was instead due to a reduction in water pressure caused by increased demand, the LA Times reported. Janisse Quiñones, the chief executive and chief engineer of the Los Angeles Department of Water and Power, told the paper that so much water was being used that the utility wasn’t able to fill the tanks quickly enough. The demand for water at lower elevations was also preventing the utility from refilling tanks at higher elevations, according to the LA Times.

Wednesday night, as powerful Santa Ana winds spread the fires to Altadena, Pasadena, and the Hollywood Hills, right-wing influencers accused city officials of slashing the fire department’s budget and prioritizing diversity programs over sound fire prevention policy — which Musk reposted on X.

But as Politico pointed out, Los Angeles mayor Karen Bass didn’t cut $23 million from the fire department’s budget, a claim that has been repeated by critics on both the right and the left, some of whom said Bass cut fire department funding to pay for a new police contract. The fire department’s budget actually increased by more than $50 million over the previous year, according to Politico, though others have noted that LA fire chief Kristin Crowley criticized Bass’ decision to cut $7 million from the department’s overtime budget just a few weeks before the Palisades fire. “The reduction … has severely limited the department’s capacity to prepare for, train for, and respond to large-scale emergencies, including wildfires,” Crowley wrote in a December 4th memo.

This nuance is, of course, being lost on X, where influencers have also begun posting AI-generated images of looters descending upon the Palisades and the Hollywood sign engulfed in flames. Disinformation is spreading like... you can probably guess what.

You may be wondering how to delete your Facebook account now that fact-checking is no longer considered important, and Meta’s changing its definition of what constitutes Hateful Conduct. It’s easy to do, and we’ll show you how. But, you should download all your stuff first.

The following instructions are for the web version of Facebook, but you can follow pretty much the same sequence on the mobile app.

Download your archives

Your Facebook archives contain just about all of the pertinent information related to your account, including your photos, active sessions, chat history, IP addresses, facial recognition data, and which ads you clicked. That’s personal information you should save.

• Click on your personal icon in the upper-right corner.
• Go to Settings & Privacy > Settings.
• Click on the Accounts Center box on the left.

Screenshot: Meta

• Go to Your information and permissions on the left, and then Download Your Information > Download or transfer information.
• You can choose to transfer information from your Facebook or Instagram account (or both).
• You now have a...

Read the full story at The Verge.

Counsel for plaintiffs in a copyright lawsuit filed against Meta allege that Meta CEO Mark Zuckerberg gave the green light to the team behind the company’s Llama AI models to use a data set of pirated ebooks and articles for training. The case, Kadrey v. Meta, is one of many against tech giants developing AI […]

© 2024 TechCrunch. All rights reserved. For personal use only.

Hello Windows Insiders, today we are releasing Windows 11 Insider Preview Build 27768 to the Canary Channel. We are also not planning to release SDKs for 27xxx series builds for the time being.

Changes and Improvements

[General]

• This update includes a small set of general improvements and fixes that improve the overall experience for Insiders running this build on their PCs.

[File Explorer]

• We’ve added a New Folder option in the context menu when right-clicking locations in the navigation pane.

[Settings]

• The ability to change time zones is available again under Settings > Time & Language > Date & Time for standard (non-admin) users.

Fixes

[File Explorer]

• Fixed an issue where the date/time properties of files were unexpectedly updated in some cases after copying them.
• Fixed an issue where File Explorer may sometimes lose focus on the search box while typing.
• Fixed an issue where when you did a search it may unexpectedly trigger the search happening repeatedly.
• Fixed an issue where the Details Pane icons weren’t responding correctly to theme changes, making it hard to see if you switched from dark to light or vice versa.
• Fixed a black flash in File Explorer that appeared when duplicating a tab.

[Input]

• Fixed an issue which could lead to the mouse cursor unexpectedly stuttering when you moved it around sometimes.

[Graphics]

• Fixed an underlying issue which could result in some games appearing oversaturated when using Auto HDR.

[Other]

• Fixed an issue causing some Insiders in Canary to experience bugchecks with error PAGE_FAULT_IN_NONPAGED_AREA.

Known issues

[General]

• [IMPORTANT NOTE FOR COPILOT+ PCs] If you are joining the Canary Channel on a new Copilot+ PC from the Dev Channel, Release Preview Channel or retail, you will lose Windows Hello pin and biometrics to sign into your PC with error 0xd0000225 and error message “Something went wrong, and your PIN isn’t available”. You should be able to re-create your PIN by clicking “Set up my PIN”.
• We’re working on the fix for an underlying issue causing accent colored window borders to not be not displayed when enabled, shadows around windows not displaying when enabled, and window launching (and other) animations to show even though the setting to show animations is turned off.
• [NEW] We’re working on the fix for an issue where SFC /scannow is showing errors every time it’s run.

[File Explorer]

• [NEW] If you open a File Explorer window and minimize it in this build, it may not render correctly when you restore it.

Reminders for Windows Insiders in the Canary Channel

• The builds we release to the Canary Channel represent the latest platform changes early in the development cycle and should not be seen as matched to any specific release of Windows and features and experiences included in these builds may never get released as we try out different concepts and get feedback. Features may change over time, be removed, or replaced and never get released beyond Windows Insiders. Some of these features and experiences could show up in future Windows releases when they’re ready.

• Many features in the Canary Channel are rolled out using Control Feature Rollout technology, starting with a subset of Insiders and ramping up over time as we monitor feedback to see how they land before pushing them out to everyone in this channel.
• Some features may show up in the Dev and Beta Channels first before showing up in the Canary Channel.
• Some features in active development we preview with Windows Insiders may not be fully localized and localization will happen over time as features are finalized. As you see issues with localization in your language, please report those issues to us via Feedback Hub.
• To get off the Canary Channel, a clean install of Windows 11 will be required. As a reminder - Insiders can’t switch to a channel that is receiving builds with lower build numbers without doing a clean installation of Windows 11 due to technical setup requirements.

• The desktop watermark shown at the lower right corner of the desktop is normal for these pre-release builds.
• Check out Flight Hub for a complete look at what build is in which Insider channel.

Thanks, Amanda & Brandon

The Azure OpenAI Assistants API introduced powerful features such as:

• Conversation State Management: Efficiently handling token limits and state.
• Automated Tool Actions: Automatically interpreting user inputs to execute actions, such as writing Python code (Code interpreter) or performing vector searches.
• Function Calling: Dynamically identifying and suggesting appropriate custom functions for execution.

Azure AI Agent Service enhances these capabilities by integrating advanced tools and actions, streamlining the developer experience even further:

1. REST APIs: Supports REST APIs compliant with Open API 3.0. It automatically identifies and invokes the appropriate API based on user intent using Swagger definitions.
2. Azure Function Apps: Azure Function Apps can be automatically invoked based on user intent, as with REST APIs mentioned above.
3. Knowledge Tools: Includes Bing Search and Azure AI Search, for data grounding.
4. Azure Logic Apps: Tooling support for Azure Logic Apps is expected but not available at the time of this writing. Currently, function calling must be used to identify and invoke the appropriate Logic App manually.
5. Choice of Language Models: Apart from gpt-4o, gpt-4o-mini models, it also supports the use of Llama 3.1-70B-instruct, Mistral-large-2407, Cohere command R+

Working with the Azure AI Agent Service

I have explored the Azure AI Agent Service’s capabilities through the sample Contoso Retail Shopping Assistant, a simple conversational AI bot.

It implements a combination of tool actions, mentioned below

• REST API Integration: Users can search for orders by category, by category and price, or place orders. Using Swagger definitions, the Azure AI Agent service dynamically identifies and calls APIs based on user input without custom code.
• Azure Logic Apps: Handles shipment order creation post-purchase. Manual function calling is used for creating shipment orders.

Apart from that, the App uses:

• Natural Language Processing: Leverages the gpt-4o-mini model to understand and process user queries effectively.
• Microsoft Bot Framework is used to build the sample App. 

Creating the Agent

Create the AI Agent with access to the necessary tools.

1. OpenApiTool - for REST API invocation
2. FunctionTool - for function calling (to trigger Azure Logic App)

functions = FunctionTool(functions=user_functions) def create_agent(): project_client = AIProjectClient.from_connection_string( credential=DefaultAzureCredential(), conn_str=config.az_agentic_ai_service_connection_string, ) # read the swagger file for the Contoso Retail Fashion API definition with open("./data-files/swagger.json", "r") as f: openapi_spec = jsonref.loads(f.read()) auth = OpenApiAnonymousAuthDetails() # Initialize agent OpenApi tool using the read in OpenAPI spec api_tool = OpenApiTool( name="contoso_retail_fashion_api", spec=openapi_spec, description="help users order a product based on id and quantity, search products by category, and search for products based on their category and price.", auth=auth, ) # Initialize agent toolset with user functions toolset = ToolSet() toolset.add(functions) toolset.add(api_tool) # print("toolsets definition", toolset.definitions) agent = project_client.agents.create_agent( model="gpt-4o-mini", name="contoso-retail-fashions-ai-agent", instructions="You are an AI Assistant tasked with helping the customers of Contoso retail fashions with their shopping requirements. You have access to the APIs in contoso_retail_fashion_api that you need to call to respond to the user queriest", tools=toolset.definitions, ) print(f"created agent with id {agent.id}")

 

The function that makes the call to Azure Logic Apps to create the delivery order:

 

def create_delivery_order(order_id: str, destination: str) -> str: """ creates a consignment delivery order (i.e. a shipment order) for the given order_id and destination location :param order_id (str): The order number of the purchase made by the user. :param destination (str): The location where the order is to be delivered. :return: generated delivery order number. :rtype: Any """ api_url = config.az_logic_app_url print("making the Logic app call.................") # make a HTTP POST API call with json payload response = requests.post( api_url, json={"order_id": order_id, "destination": destination}, headers={"Content-Type": "application/json"}, ) print("response from logic app", response.text) return json.dumps(response.text)

For simplicity, the HTTP Callable endpoint from the Azure Portal is used directly, complete with the SAS Token. The function 'create_delivery_order' calls this URL directly to trigger the Logic App.

The agent needs to be created only once. The id of the created agent must be set in the '.env' config of the Bot App.

Using the Agent in the Bot App

Retrieve the agent created in the previous step, based on the agent id.

# Create Azure OpenAI client self.project_client = AIProjectClient.from_connection_string( credential=DefaultAzureCredential(), conn_str=self.config.az_agentic_ai_service_connection_string, ) # retrieve the agent already created self.agent = self.project_client.agents.get_agent(DefaultConfig.az_assistant_id) print("retrieved agent with id ", self.agent.id)

 

Create a thread for the User Session, which the Bot App persists across conversation exchanges with the user. Every message from the user during the session gets added to this thread, and 'run' called to execute the user request.

l_thread = conversation_data.thread if l_thread is None: # Create a thread conversation_data.thread = self.project_client.agents.create_thread() l_thread = conversation_data.thread # Create message to thread message = self.project_client.agents.create_message( thread_id=l_thread.id, role="user", content=turn_context.activity.text ) run = self.project_client.agents.create_run( thread_id=l_thread.id, assistant_id=self.agent.id ) print(f"Created thread run, ID: {run.id}") while run.status in ["queued", "in_progress", "requires_action"]: time.sleep(1) run = self.project_client.agents.get_run( thread_id=l_thread.id, run_id=run.id ) if run.status == "requires_action" and isinstance( run.required_action, SubmitToolOutputsAction ): print("Run requires function call to be done..") tool_calls = run.required_action.submit_tool_outputs.tool_calls if not tool_calls: print("No tool calls provided - cancelling run") self.project_client.agents.cancel_run( thread_id=l_thread.id, run_id=run.id ) break tool_outputs = [] for tool_call in tool_calls: if isinstance(tool_call, RequiredFunctionToolCall): try: print(f"Executing tool call: {tool_call}") output = functions.execute(tool_call) tool_outputs.append( ToolOutput( tool_call_id=tool_call.id, output=output, ) ) except Exception as e: print(f"Error executing tool_call {tool_call.id}: {e}") print(f"Tool outputs: {tool_outputs}") if tool_outputs: self.project_client.agents.submit_tool_outputs_to_run( thread_id=l_thread.id, run_id=run.id, tool_outputs=tool_outputs, ) print(f"Current run status: {run.status}")

Note that there is no code to be written to make the REST API calls. The OpenApiTool calls the appropriate API automatically during the thread run. Only in the scenario where the delivery order needs to be created, the execution flows into the function calling block, where the identified function is manually executed.

 

The source code ofthe Bot App discussed in this post is available here

See a demo

Watch a demo of the sample app, below.

 

Additional Resources

For more details, check out:

• Azure AI Agent Service documentation.
• Microsoft Bot Framework Documentation
• Azure Logic Apps Documentation

Learn how to remove Azure Functions and their dependencies from Microsoft Teams Toolkit React projects for simpler, more efficient tab applications.

Read the full article (11 minutes reading time): How To Remove Azure Functions from React Tab TTK Projects.