CONSUMER

Why Apple enthusiasts should try Siri AI now

Apple's public betas for the new software that will power iPhone, Mac, and iPad have arrived, and that means more people can try out the new version of Siri. 

Spoiler: The new Siri AI is better, and it's worth trying if you're an Apple enthusiast. Just know the caveats before you install it on everyday devices.

On Monday, Apple released the public beta versions of iOS 27, macOS 27, iPadOS 27, and watchOS 27. That means anyone can download and test Siri AI to see for themselves if it lives up to being a more useful AI assistant. 

The developer betas have been out since the new software was announced on June 8 at WWDC, and they have already advanced to beta version 3. The final version of the software won't come until this fall, typically late September, around the time the new iPhone ships. 

I've been using iOS 27, MacOS 27, and iPadOS 27 since June 8 and here are my thoughts on the new Siri AI:

• Works solidly, but it's not frontier AI: While Siri used to be enormously frustrating to use and most users would just give up on it, the new Siri AI is far more functional and reliable. It can answer basic questions and pull up information you'd expect with a lot less friction. It won't give you the level of detail or accuracy you'd get from Claude or ChatGPT, but it works now. And the ability to better access for personal information safely is its superpower.

• The visuals look great: In typical Apple fashion, the designers have outdone themselves. The Siri AI animations that show up when the model is registering your voice input, for example, both look smooth and are a very useful visual indicator. It looks better than any other app or system, and the only thing even close is ChatGPT Voice. Of course, the previous version of Siri looked pretty good, too. This one actually works, though.

• Best comparison is Meta: In terms of the usefulness and efficiency of Siri AI, I'd compare it to the latest version of Meta smartglasses running the new Muse Spark models. If you have Meta glasses and haven't tried that recently, give it a go. It's also a big upgrade. Both of these products are more useful now, and their level of detail and accuracy is pretty similar. They aren't going to rival OpenAI, Anthropic, or Google. But they are good enough for quick stuff now.

The public betas are fairly stable, and I have no problem recommending Apple enthusiasts install them if you want to try out the new Siri AI. Just make sure you run a full backup of your device before you install the beta. Also, if you run your business on the device or need it to be fully functional for your personal safety, do not install the public beta on it. If you are a serious Apple enthusiast and have a more recent device from the past few years that you've replaced recently and aren't using, then install the beta on that device. That's what I did on iPhone and Mac. On iPad, I didn't mind installing the beta software because I don't use it for anything mission-critical — just reading and watching videos. Once you give the new Siri AI a try, let me know how it's going for the ways you use it by pinging me on Twitter.

TOGETHER WITH CONVEX

Code Faster, Easier, And Less Buggy With Convex

You’re probably already familiar with Convex, the backend building platform that helps your AI agents excel – but their latest updates are taking things up a notch. Convex recently launched plugins that allow you to connect Claude Code, Codex, GitHub Copilot and more directly to your Convex deployment, allowing them to…

• Read data and run functions entirely on their own

• Optimize your project’s performance via production insights

• Improve safety by limiting access to production

There’s even more updates coming in the near future, but for now, see how you can take your agents’ work to the next level with the latest from Convex right here.

GOVERNANCE

Why 16 Nobel laureates raised the red flag on AI

AI's risks have gone beyond speculation. 

Nearly 200 economists and tech leaders have signed a new petition titled "We Must Act Now," highlighting the negative impacts AI technology will have on the economy, including job displacement, and urging leaders to take action now. The signees include prominent figures such as the chief economists of OpenAI and Anthropic, as well as 16 Nobel laureates. 

The statement, available for anyone to sign, is a concise, three-point list that summarizes the potential impact and concludes with a call to action: 

1. Timeline: "AI may become radically more powerful over the next 10 years."

2. Impact: "This could drive an unprecedented transformation of our economy, larger than the Industrial Revolution, but unfolding over a vastly shorter time frame. It could bring risks, including large-scale job displacement, as well as opportunities such as major gains in living standards."

3. Call-to-action: "Economists, policymakers and technology leaders must act now to understand the economics of transformative AI and to build the incentives, guardrails, and institutions needed to steer AI in a direction that complements humans and benefits society."

One of the petition's signers, Alvin Wang Graylin, digital fellow at the Stanford Institute for Human-Centered Artificial Intelligence (HAI), told The Deep View, "AI has become so polarizing these days that the noise is drowning out the very real problems right on our doorstep. If we keep debating instead of building, testing, and deploying solutions, the crises will arrive well before the benefits do. The time to act is now to prevent problems, not after the disruption hits, as traditionally happens."

Petitions highlighting concerns about the reach of AI, given its rapid development, are nothing new. One of the most notable is a 2023 petition with over 1,800 signatories, including Elon Musk and Apple co-founder Steve Wozniak. However, the biggest difference here is that this focuses on the economic impacts, which have been highly debated among experts. 

For instance, Preston Caldwell, the senior US economist for Morningstar Research Services, says early evidence signals that AI-driven automation may be no different from the automation that has occurred with other technologies throughout US history, including manufacturing and agriculture. 

“Adjusted for benchmark performance, the price of large language model queries has plummeted by orders of magnitude over the past few years. If AI becomes cheap enough, then human workers can still claim a large share of GDP even while focusing on a progressively smaller range of non-automatable tasks," said Caldwell.

Of course, this comes against the backdrop of a series of layoffs that executives have pinned on AI. This year alone, Oracle, Meta, Cisco, Cloudflare, Coinbase, and more have blamed layoffs on AI. Founders Reports has put together an AI layoff tracker and claims that through June 2026, AI has been cited in 101,743 U.S. job cuts, nearly double the 54,836 attributed to it in all of 2025.

There's no denying that the speed at which AI is advancing and being deployed is unprecedented. The result is that no one truly understands what the outcome will be, and when something this disruptive touches the economy and people's livelihoods, that uncertainty demands care and scrutiny. That's something everyone across the board can agree on, regardless of where you land on the actual solutions. Even in the best-case scenario, where the AI boom doesn't turn to bust and instead becomes a durable part of the US economy, there will likely be a painful transition period, just as any structural shift produces short-term growing pains before things stabilize. And given AI's sheer reach, that short-term pain could still mean real disruption for a substantial portion of the population.

TOGETHER WITH CRUSOE

Tune it. Deploy it. Own it.

Crusoe Serverless Fine-Tuning is now generally available in Crusoe Intelligence Foundry. Customize leading open models — Qwen, DeepSeek, Gemma, gpt-oss, and more — on your proprietary data without provisioning a single cluster.

Token-based pricing tracks spend directly to model progress, and early stopping ends billing the moment your model stops improving. Deploy in one click with Self-Serve Deployments, or download your weights in .safetensors format and run them anywhere. No infrastructure wrangling. No surprise bills. Just breakthroughs.

Start fine-tuning

PRODUCTS

Why the summer of AI price wars is here

More powerful AI models have typically carried higher usage limits due to their greater compute demand. That is no longer the case.

In June, Anthropic made its Claude Fable 5 available to users, a model so capable that it is described as "Mythos-class," a term used to refer to the Mythos model, deemed so powerful that it cannot be widely released. Then, just last week, OpenAI released its counterpart, GPT-5.6. The capabilities of these models show a real step change in performance compared to their predecessors—perhaps more than ever before. 

Since the models require as much compute as they do, they were not offered to all users for free at launch. Rather, both companies made it available to their paid subscribers with slightly different approaches: Anthropic made Fable available for a certain amount of time, originally until July 7, while OpenAI made its mid-tier model, GPT-5.6 Terra, available to free users, reserving the flagship Sol and the lightweight Luna for paid, higher plans with some usage limits.

Sensing the pressure from OpenAI's release, Anthropic extended its promotional access to the Fable 5 model for all paid subscribers through July 19 and kept weekly rate limits for Claude Code 50% higher. Thibault Sottiaux, engineering lead for Codex, responded to the Anthropic announcement, acknowledging the competition and saying, "I think GPT 5.6 is pretty good." 

Adding to the competition, OpenAI celebrated its seven million active users of Codex and ChatGPT by adding a banked reset to everyone's account. This lets you reset your usage caps if you hit the limit from doing a lot of work with the models. 

Again, the comments were flooded with users celebrating the extension and comparing OpenAI's release to Anthropic's. 

The reasoning for this back-and-forth is an effort to capture the business of power users. There is clearly user demand, as evidenced by the need to reset quotas and by the X threads celebrating each limit extension in the announcement post. The Deep View's Jason Hiner ran through his quota this weekend and said he could've used the reset, because he ended up having to buy extra tokens even though he's on the Pro plan. 

Though they could be making more money in the short run by charging users, the idea is to keep users using the AI even more, which is especially valuable for AI tools because in that time users are building their memory, and once that is built, users are more likely to stay within the same AI since it's gotten to know them. Same for any agents built, which may be difficult to transfer.

Of course, there are other self-interested factors these companies have to appease users, as Rob May, CEO and co-founder of Neurometric, told The Deep View.

"Anthropic had some outages in the past week and this was probably a response to that to make users happy.  They needed a mea culpa for the outages and resetting the usage limits was the easiest way," said May. "OpenAI does it to drive adoption of their models for coding usage because that's where Anthropic is winning, and the technical differentiation between the two models isn't really there."

Competition is also coming from other companies, with underdogs such as Meta and SpaceXAI continuing to release their own models focused on reducing token costs. Ultimately, users are the beneficiaries of this competition. They can try out both models and then decide which one they might want to commit to as a subscription. According to Morningstar research, "the US AI sector is already yielding massive revenue from end users, with perhaps $100 billion in AI services revenue in 2025."

Ultimately, each AI company is betting that if it gives users free access to some of its top models, those users will become hooked and keep coming back for more. This is a strategy businesses have used since the beginning of time, like when Costco took a loss on its $1.50 hot dog, betting that the loss leader would drive people into the store to shop. But there is an important nuance here: compute and tokens are extremely finite resources right now. Even if they sound like abstract concepts, every free query still carries a real cost. Whether this bet actually pays off is something we can only measure over the long run, and it will likely become much clearer once these companies IPO, and we can actually analyze their earnings reports. But in the end, it will likely depend on how rapidly AI's unit economics decline. 

LINKS

• Meta committed to spending another $40 billion on its data center campus in Louisiana

• Intel to invest €5 billion in Irish chip manufacturing facility

• Samsung moves up Yongin chip plant opening to 2029 from 2030/2031 timelines

• Anthropic hired Monzo cofounder Tom Blomfield, one of the biggest names in UK tech

• Holiday Robotics raised $103 Million Series A; the largest in Korean startup history

• OpenAI, Meta, SpaceXAI, race to release cost-efficient models

• Perplexity: The app now has an account switcher 

• Cerebras: Gemma 4 is now available on Cerebras

• Hy3: Claimed #1 on the OpenRouter LLM leaderboard

• Blume: "A world-class docs framework for everything you ship"

• Codex/ChatGPT Work: To celebrate 7M active users, OpenAI added a banked reset to everyone's account

• Within: AI Engineer 

• Nobert Health: Applied AI Engineer

• Blossom: Applied AI Software Engineer

• OpenAI: Applied AI Engineer, Codex Core Agent

GAMES

POLL RESULTS

Will you be trying any of OpenAI's new GPT-5.6 models?

Yes (40%)
No (31%)
Maybe (29%)