The-Claude-lash-is-here-Opus-47-is-burning-through-tokens.jpeg

The Claude-lash is here: Opus 4.7 is burning through tokens — and some people’s patience

by with No Comment Xpandfy

Anthropic says its new AI model, Opus 4.7, should feel “more intelligent, agentic, and precise.” Some users aren’t feeling the joy.

The backlash to the new Claude model — a relatively rare occurrence for an AI product many view as the gold standard in AI — has gained traction on social media.

Anthropic is coming off months of widespread public celebration. The technical chops of Claude Code and Claude Cowork have boosted the company’s image, and chatbot fans have long admired Claude’s writing abilities. After the company’s fight with the Department of Defense, Claude went No. 1 in the App Store.

But hot on the heels of Anthropic users fretting that the previous model, Opus 4.6, had been “nerfed,” the early reactions to 4.7 indicate Anthropic has a growing Claude backlash on its hands.

What are people saying about Opus 4.7?

There are examples of supposed Opus 4.7 flubs across social media.

One Reddit post titled, “Claude Opus 4.7 is a serious regression, not an upgrade,” has 2,300 upvotes. An X user‘s suggestion that Opus 4.7 wasn’t really an improvement over Opus 4.6 got 14,000 likes.

In one informal but popular test of AI intelligence, Opus 4.7 appears to say that there were two Ps in “strawberry.” Another user screenshot shows it saying that it didn’t cross reference because it was “being lazy.” Some Redditors found that Opus 4.7 was rewriting their résumés with new schools and last names. Multiple X users posited that Opus 4.7 had simply gotten dumber.

Some X users have suggested the culprit is the AI model’s reasoning times. Anthropic says the new “adaptive reasoning” function lets the model decide when to think for longer or shorter periods. One user wrote that they couldn’t “get Opus 4.7 to think.” Another wrote that it “nerfs performance.”

“Not accurate,” Anthropic’s Boris Cherny, the creator of Claude Code, responded. “Adaptive thinking lets the model decide when to think, which performs better.”

In other cases, Anthropic has acknowledged there’s room for improvement.

After one user specifically flagged issues with the adaptive reasoning on the Claude website, an Anthropic product manager responded that the team was “sprinting on tuning this more internally and should have some updates here shortly.”

Gergely Orosz, the writer of the “Pragmatic Engineer” newsletter, posted a screenshot of Claude not knowing what OpenClaw was. Cherny chimed in to ask whether he had “web search” enabled. He hadn’t, it turned out, but Orosz wrote that he had never “touched it” before.

Orosz also said he found the model “surprisingly combative.” (He “gave up” and returned to Opus 4.6.) Others found that the model refused to code certain prompts or put up safety flags on simple images.

Opus 4.7 can also use up a whole lot more tokens. The model has a new tokenizer, which means that one input can cost roughly 1.0–1.35x as many tokens as it would with a previous model.

After its release, one X user said that Claude Pro subscribers could ask only three questions before hitting their limit. Another user noticed that Opus 4.7 was priced at a 7.5x premium in GitHub Copilot until the end of April. “Yeeaaah I’ll stick to 4.6 for now,” they wrote.

Cherny later announced that Anthropic was increasing subscriber rate limits “to make up for it.”

Those upset with Opus 4.7 may look back to an old model like 4.5 — only to find that it’s gone. One Reddit thread is full of self-described “heartbroken” and “grieving” fans of 4.5.

It’s not unusual for AI companies to face user pushback when depreciating an older model that proved popular — look no further than when OpenAI took away GPT-4o.

As was the case with OpenAI’s popular 4o model, some fans have turned to trying to bargain with Anthropic.

“Please open back support for Opus 4.5,” one Redditor wrote under Anthropic’s post. “4.6 is unusable and 4.7 eats usage like nuclear reactor.”

Anthropic did not respond to a request for comment from Business Insider.

Do people like Opus 4.7?

Not everyone is put off by the resulting pricing. “Opus 4.7 is burning through tokens like nobody’s business, but it’s gooooooooood,” one X user wrote.

Anthropic says that 4.7 is “a notable improvement on Opus 4.6 in advanced software engineering, with particular gains on the most difficult tasks.”

“Users report being able to hand off their hardest coding work—the kind that previously needed close supervision—to Opus 4.7 with confidence,” the AI company wrote in its announcement.

While some users threaten to leave for OpenAI’s models, others are singing Opus 4.7’s praises. Startup founder Jeremy Howard described it as “the first model that ‘gets’ what I’m doing when I’m working.”

Y Combinator CEO Garry Tan wrote that he’s using it for his OpenClaw, and Cursor designer Ryo Lu said he uses it for planning.

For those still skeptical of Opus 4.7, Anthropic seems to be tinkering amid the feedback.

“A lot of bugs that folks may have hit yesterday when first trying Opus 4.7 are now fixed,” Anthropic staffer Alex Albert wrote on Friday. “Thanks for bearing with us.”




Source link

A-Google-engineer-whose-job-is-changing-due-to-AI.jpeg

A Google engineer whose job is changing due to AI explains how she’s learning without burning out

by with No Comment Xpandfy

This as-told-to essay is based on a conversation with Pratiksha Patnaik, a 30-year-old cloud infrastructure engineer at Google Cloud Consulting, based in Seattle. Her identity and employment have been verified by Business Insider. The following has been edited for length and clarity.

I’ve been with Google for around three years and I started as an infrastructure engineer. I’m still an infrastructure engineer and on a day-to-day basis, I work with customers to build different solutions, depending on the needs of the customer.

At first, I was mostly involved with networking security and infrastructure customers. But as we saw the AI wave come in, we started focusing more on customers that want to adopt gen AI products and solutions.

I didn’t transition into an AI role, but I’m working with a lot of AI services, and AI engineers who are working on features for those services. My job is a combination of working with customers and the product team, to provide technical solutions for customers. It’s a constant feedback loop to figure out if the solution we’re building is right for the customer we’re working for.

Our job is to know how these products work. Sometimes when we work on the products, we identify feature gaps or bugs, so we need to work with the product team or engineering team.

I’ve been in the same role the whole time, but the nature of my job is changing because of everything going on in the AI space. We get a lot of demand in AI products and we have to do a lot of trainings on it to deliver.

I spend an hour or two weekly on trainings

The more AI progresses, the more difficult it’s become to keep up. As the rate of AI innovation gets quicker, the role of engineers has transitioned from mastery to continuous adaptation at scale.

Just being aware of everything that’s happening in the tech industry, along with what we have to do with the customer, has changed dramatically from what it was like a year ago. Back then, we had to execute within known constraints. But as time passes and AI evolves rapidly, those perimeters have dissolved and we have to invest much more time to learning about changes in this space. We now have to navigate an ever-expanding problem space alongside our customers.

I spend around one to two hours a week up-skilling on new concepts. We have a lot of internal trainings that we can utilize. So I see if there is something new that I’m interested in learning about and that can help me do my job.

I am gaining a deeper level of understanding in high-performance computing, AI observability, model performance benchmarking, and the underlying architecture of GPUs and TPUs.

It can get overwhelming

The culture at Google is very much about constantly learning. Every day we learn about a new tool or model version. That motivates me to keep learning. We also have to skill up in order to put our best foot forward in front of the customers.

But with the pace of technology nowadays, I feel like I need to know everything — and if I don’t learn, I might be left behind.

The reality is that it’s not practically possible to know everything with the changes that are coming out at an exponential rate. To remain effective without burning out, I prioritize intentional depth over exhaustive consumption. By focusing on what really interests me, I can make sure that my learning is not just a chore of “keeping up,” but an investment in expertise.

When I read too much, I get overwhelmed and it’s not possible to retain all of the information I’m consuming. We’re at a point where the amount of information we have is huge and we have to figure out where to spend our time and what’s the most beneficial for us.

Are you an engineer experiencing changes in your job? We’d love to hear from you. Email the reporter from a non-work device via email at aaltchek@insider.com or secure-messaging platform Signal at aalt.19.




Source link

Satellite-images-show-an-Iranian-warship-burning-pierside-after-US.jpeg

Satellite images show an Iranian warship burning pierside after US and Israeli strikes

by with No Comment Xpandfy

New satellite imagery shows an Iranian warship burning pierside following widespread US and Israeli airstrikes.

The imagery, collected Saturday by US spatial intelligence firm Vantor and obtained by Business Insider, shows some of the initial aftermath of the joint US and Israeli attack on Iran, a major combat operation against Tehran after the collapse of talks on nuclear weapons.

The Vantor images show a frigate burning at the Konarak naval base in southern Iran. A large billow of smoke can be seen rising from the wounded vessel.


Overview of a burning vessel at Iran's Konarak naval base on February 28.

A frigate on fire at Iran’s Konarak naval base.

Satellite image ©2026 Vantor.



In a close-up image, two other frigates are seen sitting nearby, appearing unscathed.

Trump said earlier in his video message announcing the US campaign, Operation Epic Fury, that the US military was “going to annihilate” the Iranian navy, as well as the country’s other military capabilities.

Trump said the goal was to ensure that Iran can “never” have a nuclear weapon. The president also urged Iranian citizens to “take control” in the aftermath of US operations.

“This will be, probably, your only chance for generations,” he said.


A burning frigate at Konarak naval base in Iran.

A close-up of a burning frigate at Konarak.

Satellite image ©2026 Vantor.



War analysts at the Institute for the Study of War think tank said on Saturday that there had been unconfirmed reports of strikes on Iranian naval assets.

“An Israeli OSINT [open source intelligence] account reported strikes on the IRGC Navy frigate Jamaran,” ISW said. “There are also unverified reports of strikes on the IRGC Navy Imam Ali Navy Base in Chabahar, Sistan and Balochistan Province.”

Other images collected by Vantor show drone launch activity at two nearby airbases in Konarak. After the initial attacks by the US and Israel, Iran launched retaliatory missile and drone strikes at US military bases around the region in Bahrain, Qatar, the United Arab Emirates, Kuwait, Jordan, and Saudi Arabia.


Drones dispersed at an airbase in Iran on February 28.

Drones dispersed at an airbase in Konarak.

Satellite image ©2026 Vantor.



Iran has also launched multiple barrages at Israel, including missiles and drones, and said it was carrying out an “extensive drone offensive operation” against its longtime foe.

US strikes on Iran included a mix of weapons such as Tomahawk cruise missiles, High Mobility Artillery Rocket Systems, drones, and undisclosed standoff weapons designed for long-range strikes outside the reach of enemy air defenses.

In response to retaliatory attacks from Iran, various air defense weapons have been active, including the MIM-104 Patriot surface-to-air missile and Terminal High Altitude Area Defense, or THAAD, batteries.


Drones dispersed and a runway blocked at the Konarak drone strip.

Drones dispersed on a blocked runway.

Satellite image ©2026 Vantor.



Saturday’s strike is the second time in less than a year that the US has launched a strike campaign against Iran. In June 2025, US forces targeted Iranian nuclear facilities. In the weeks prior to the most recent attack, the US built up a massive force presence in and around the Middle East.


Source link