• Tech employees share their salary negotiation tips, which helped boost their pay by tens of thousands of dollars.
• Their negotiation strategies include practicing pitches, using data, and leveraging multiple offers.
• Research and transparency are key in negotiating better compensation in tech roles, they said.

Sarra Bounouh has worked at consulting giant Accenture and three Big Tech companies.

But she still deals with imposter syndrome, especially when talking compensation.

"Going into a negotiation is always, at least for me, a very uncomfortable discussion," Bounouh told Business Insider. "I just want to push through and ask for what I deserve."

She and four other tech employees from Meta, Google, and Cisco shared their salary negotiation tips before joining a company or when trying to get promoted. They have used these strategies to add tens of thousands of dollars to their original offers in recent years.

Product manager at Meta

Avoid offering the first number. If you must, back it up with research, said Bounouh, a product manager who joined Meta earlier this year.

She suggested using resources like Levels.fyi or Glassdoor and selecting your role and geography to see recent offers and compensation that makes sense for that job.

"I personally don't like having detailed conversations about level and compensation from that first call with the recruiter because I want to meet the team, I want to meet the hiring manager, I want to get excited about the role," she said.

Bounouh prefers to negotiate her level and compensation once there's an offer on the table.

She said she often gets asked about salary expectations early in the process because recruiters say they want to save time for both sides.

She politely declines to share a number by telling the recruiter: "I don't have a number for your right now. I will need to do some research before getting back to you. At this stage of the process, I'm more focused on meeting the hiring manager and team."

Rehearsal is key for conversations about promotions or raises, she said.

Bounouh said she practiced her pitch for every job after Accenture and increased all three jobs' initial salary offers: Microsoft by 32%, Snap by 19%, and Meta by 37%.

Product manager at Oracle

Internal transfers between teams or offices are also an opportunity to negotiate your compensation package.

Ketaki Vaidya, who moved from Oracle's India to California office in 2022, said she approached her negotiation with an "everything under the sun is negotiable" mindset.

First, Vaidya looked at Glassdoor and talked to people who'd made the move to gather salary data. She wanted to ensure she was getting a fair offer for the US' cost of living.

"I was being given this offer for the credibility that I had built in the organization. I felt like I had an upper hand in negotiating," she said. "I was much more confident in asking for the things that I deserve — so it ended up being a very smooth transition."

After negotiating her base salary up to $80,000, she discussed other compensation components, including the timing of her next review, sign-on bonuses, relocation costs, paid leave, and remote work. She negotiated a sign-on bonus of $15,000 and a relocation allowance of $15,000, which weren't part of the initial offer.

Now, her compensation is about $130,000 annually, including stock units and bonuses.

Product manager at Cisco

When Varun Kulkarni switched from consulting to tech to work on more artificial intelligence projects, he was careful not to come off as aggressive during his pay negotiations.

Once he had offers from Cisco and others in hand in 2022, he was transparent with recruiters and mentioned other offers, without introducing his own counter number.

He asked recruiters how high they could go and what they thought about other offers.

"You want to kind of not be too pushy" he said.

His offer from Cisco already matched the market rate and what several competitors were offering, but he managed to negotiate it by 5%, bringing his total compensation to $180,000.

Product manager at Google

During his 2022 recruitment process at Google, Yung-Yu Lin used his employer at the time, PayPal, to land better offers from both companies.

He interviewed and landed jobs at several places — but their pay did not compare with Google's offer.

Lin decided to negotiate a retention package. PayPal countered with a 10% pay bump. He then renegotiated with Google.

Google offered a 20% raise on his original compensation at PayPal, which brought his offer to the $350,000 to $400,000 range as a senior product manager, including stock-based compensation.

Software engineer at Meta

Hemant Pandey, a senior software engineer at Meta, used other offers and research in his most recent job search.

After two years at Salesforce, in 2021 he applied to Meta, TikTok, LinkedIn, and two other companies. He used offers from these companies to negotiate his compensation at Meta.

"Be very transparent that you have other offers. Even if you have interviews going on, mention those, because it's also leverage," he said. It signals to the recruiter that they have to move fast and work with your parameters.

Meta's recruiters matched the base salary and restricted stock units from the highest of all offers.

Aside from being transparent, Pandey said it is important to be proactive and research how compensation works in different companies. For example, candidates should compare how stocks are refreshed, he said. A refresher is when the stock option portion of an employee's compensation is updated.

"I also negotiated my sign-on bonus and said, 'Hey, at Salesforce, I'll be leaving my $30,000 to $40,000 of annual bonus if I join you. Can you help me accommodate that?'"

Pandey was offered $520,000 in annual pay, including stock options, in that 2021 move.

"The most significant thing happened in my career when I made the move from Salesforce to Meta, which was close to almost 80 to 90% hike" in pay, Pandey said.

Do you work in tech, consulting, or finance and have a story to share about your career journey? Please reach out at [email protected].

Welcome back to Week in Review. This week, we’re looking at OpenAI’s last — and biggest — announcement from its “12 Days of OpenAI” event; Apple’s potential entrance into the foldable market; and why Databricks is choosing to wait to go public. Let’s get into it. P.S. We’re off for the holidays! Week in Review […]

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

Google has offered up its own proposal in a recent antitrust case that saw the US Department of Justice argue that Google must sell its Chrome browser. US District Court Judge Amit Mehta ruled in August that Google had acted illegally to maintain a monopoly in online search, with the DOJ then proposing a number […]

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

• Google on Friday proposed possible remedies to resolve an antitrust case over its search business.
• Last month, the DOJ suggested that the judge force Google to sell its Chrome browser.
• Judge Amit Mehta is expected to rule on the final remedies by August 2025.

Google on Friday proposed limitations to its search partnerships as a potential remedy to resolve antitrust violations in its search business.

The proposal would allow Google to continue partnering with third-party companies like Apple in revenue-sharing deals that make Google the default search engine on their devices, unlike the Justice Department's proposal. However, Google's proposal would make the deals non-exclusive, the company said in its filing.

"We don't propose these changes lightly," Google said in a blog post about the proposal. "They would come at a cost to our partners by regulating how they must go about picking the best search engine for their customers. And they would impose burdensome restrictions and oversight over contracts that have reduced prices for devices and supported innovation in rival browsers, both of which have been good for consumers."

Last month, the Justice Department and a group of states asked Judge Amit Mehta to force Google to sell its Chrome browser to resolve the case. They also asked that Google be stopped from entering default search agreements with Apple and other companies and that Google should open its search engine results to competitors.

Industry experts previously told Business Insider that selling Chrome off would open up the browser market and would likely be cheered on by search rivals and advertisers, though it remains unclear how a possible Chrome spinoff might work.

Both sides will present arguments for their proposals at a hearing scheduled for April. The judge is expected to rule on the final remedies by August.

Kent Walker, Google's president of global affairs, previously said the company intends to appeal the judge's ruling, potentially delaying a final decision by several years.

Representatives for the Justice Department's antitrust division did not immediately respond to a request for comment from Business Insider.

• OpenAI's o1 model was hailed as a breakthrough in September.
• By November, a Chinese AI lab had released a similar model called DeepSeek.
• On Thursday, Google came out with a challenger called Gemini 2.0 Flash Thinking.

In September, OpenAI unveiled a radically new type of AI model called o1. In a matter of months, rivals introduced similar offerings.

On Thursday, Google released Gemini 2.0 Flash Thinking, which uses reasoning techniques that look a lot like o1.

Even before that, in November, a Chinese company announced DeepSeek, an AI model that breaks challenging questions down into more manageable tasks like OpenAI's o1 does.

This is the latest example of a crowded AI frontier where pricey innovations are swiftly matched, making it harder to stand out.

"It's amazing how quickly AI model improvements get commoditized," Rahul Sonwalkar, CEO of the startup Julius AI, said. "Companies spend massive amounts building these new models, and within a few months they become a commodity."

The proliferation of multiple AI models with similar capabilities could make it difficult to justify charging high prices to use these tools. The price of accessing AI models has indeed plunged in the past year or so.

That, in turn, could raise questions about whether it's worth spending hundreds of millions of dollars, or even billions, to build the next top AI model.

September is a lifetime ago in the AI industry

When OpenAI previewed its o1 model in September, the product was hailed as a breakthrough. It uses a new approach called inference-time compute to answer more challenging questions.

It does this by slicing queries into more digestible tasks and turning each of these stages into a new prompt that the model tackles. Each step requires running a new request, which is known as the inference stage in AI.

This produces a chain of thought or chain of reasoning in which each part of the problem is answered, and the model doesn't move on to the next stage until it ultimately comes up with a full response.

The model can even backtrack and check its prior steps and correct errors, or try solutions and fail before trying something else. This is akin to how humans spend longer working through complex tasks.

DeepSeek rises

In a mere two months, o1 had a rival. On November 20, a Chinese AI company released DeepSeek.

"They were probably the first ones to reproduce o1," said Charlie Snell, an AI researcher at UC Berkeley who coauthored a Google DeepMind paper this year on inference-time compute.

He's tried DeepSeek's AI model and says it performs well on complex math problems that must be solved by thinking for longer and in stages. 

He noted that in DeepSeek's DeepThink mode, the model shows users every step of its thought process. With o1, these intermediate steps are hidden from users. 

"I've asked people at OpenAI what they think of it," Snell told BI. "They say it looks like the same thing, but they don't how DeepSeek did this so fast."

OpenAI didn't respond to a request for comment. On Friday, the startup previewed an o1 successor, called o3. Francois Chollet, a respected AI expert, called the update a "significant breakthrough."

Andrej Karpathy, an OpenAI cofounder, praised Google's new "Thinking" model for the same reasoning feature.

"The prominent and pleasant surprise here is that unlike o1 the reasoning traces of the model are shown," he wrote on X. "As a user I personally really like this because the reasoning itself is interesting to see and read — the models actively think through different possibilities, ideas, debate themselves, etc., it's part of the value add."

A DeepSeek demo

Snell shared a multistep math problem with Business Insider, which we used to test DeepSeek for ourselves: 

"Find a sequence of +, -, /, * which can be applied to the numbers 7, 3, 11, 5 to get to 24, where each of the given numbers is used exactly once."

BI put that prompt in DeepSeek's chat window on its website. The model responded initially by laying out the challenge ahead.

"Alright, so I've got this problem here: I need to use the numbers 7, 3, 11, and 5, and combine them with the operations of addition, subtraction, multiplication, and division, using each number exactly once, to get to 24," it replied. "At first glance, this seems a bit tricky, but I think with some systematic thinking, I can figure it out."

It then proceeded through multiple steps over roughly 16 pages of discussion that included mathematical calculations and equations. The model sometimes got it wrong, but it spotted this and didn't give up. Instead, it swiftly moved on to another possible solution. 

"Almost got close there with 33 / 7 * 5 ≈ 23.57, but not quite 24. Maybe I need to try a different approach," it wrote at one point. 

After a few minutes, it found the correct solution. 

"You can see it try different ideas and backtrack," Snell said in an interview on Wednesday. He highlighted this part of DeepSeek's chain of thought as particularly noteworthy:

"This is getting really time-consuming. Maybe I need to consider a different strategy," the AI model wrote. "Instead of combining two numbers at a time, perhaps I should look for a way to group them differently or use operations in a nested manner."

Then Google appears

Snell said other companies are likely working on AI models that use the same inference-time compute approach as OpenAI.

"DeepSeek does this already, so I assume others are working on this," he added on Wednesday.

The following day, Google released Gemini 2.0 Flash Thinking. Like DeepSeek, this new model shows users each step of its thought process while tackling problems. 

Jeff Dean, a Google AI veteran, shared a demo on X that showed this new model solving a physics problem and explained its reasoning steps. 

"This model is trained to use thoughts to strengthen its reasoning," Dean wrote. "We see promising results when we increase inference time computation!"

Google will soon take more steps to make AI a part of search, exposing more users to its Gemini agent, according to recent reports and app teardowns.

"AI Mode," shown at the top left of the web results page and inside the Google app, will provide an interface similar to a Gemini AI chat, according to The Information.

This tracks with a finding from Android Authority earlier this month, which noted a dedicated "AI mode" button inside an early beta of the Google app. This shortcut also appeared on Google's Android search widget, and a conversation history button was added to the Google app. Going even deeper into the app, 9to5Google found references to "aim" (AI mode) and "ai_mode" which suggest a dedicated tab in the Google app, with buttons for speaking to an AI or sending it pictures.

Read full article

Comments

Over the past month, we've seen a rapid cadence of notable AI-related announcements and releases from both Google and OpenAI, and it's been making the AI community's head spin. It has also poured fuel on the fire of the OpenAI-Google rivalry, an accelerating game of one-upmanship taking place unusually close to the Christmas holiday.

"How are people surviving with the firehose of AI updates that are coming out," wrote one user on X last Friday, which is still a hotbed of AI-related conversation. "in the last <24 hours we got gemini flash 2.0 and chatGPT with screenshare, deep research, pika 2, sora, chatGPT projects, anthropic clio, wtf it never ends."

Rumors travel quickly in the AI world, and people in the AI industry had been expecting OpenAI to ship some major products in December. Once OpenAI announced "12 days of OpenAI" earlier this month, Google jumped into gear and seemingly decided to try to one-up its rival on several counts. So far, the strategy appears to be working, but it's coming at the cost of the rest of the world being able to absorb the implications of the new releases.

Read full article

Comments

• Google's CEO said it had cut managers, directors, and VPs by 10% as part of its efficiency drive.
• The company has been boosting efficiency by reducing layers and reorganizing teams.
• Google has been facing down threats from OpenAI and other AI rivals.

Google had cut the number of top management roles by 10% in its push for efficiency, CEO Sundar Pichai told employees in an all-hands meeting on Wednesday.

Pichai said that Google had made changes over the past couple of years to simplify the company and be more efficient, according to two people who heard the remarks, who asked to remain anonymous because they're not authorized to speak to the press.

Pichai said this had included a 10% reduction in managers, directors, and vice presidents, one source said.

A Google spokesperson said that some of the roles in that 10% figure were transitioned to individual contributor roles and that some were role eliminations.

The company has been on an efficiency drive for more than two years. In September 2022, Pichai said he wanted Google to be 20% more efficient, and the following January, the company had a historic round of layoffs that saw 12,000 roles eliminated.

The efficiency push has coincided with AI rivals such as OpenAI unleashing new products that threaten Google's search business.

Google has responded by injecting generative AI features into its core businesses and launching a flurry of new AI features, such as a new AI video generator beating OpenAI's in early testing and a new set of Gemini models, including a "reasoning" model that shows its thought process.

In Wednesday's all-hands meeting, Pichai also clarified the meaning of the word "Googleyness," telling staff that it needed updating for a modern Google.

Are you a current or former Google employee with something to share? You can reach the reporter Hugh Langley via the encrypted messaging app Signal (+1-628-228-1836) or email ([email protected]).

• In an all-hands, Google CEO Sundar Pichai said the word 'Googleyness' had become too broad.
• Pichai clarified what the word means for the company.
• Now it's about being "Mission First" and being "Bold and Responsible."

"Googleyness" has long been a vague word for the search giant. Once used to determine if a candidate is a good fit for hiring, it has evolved in definition over the years.

Google CEO Sundar Pichai just attempted to clarify what the word means for Googlers now.

In a company all-hands meeting on Wednesday, Pichai told staff the definition of "Googleyness" had become too broad and that he felt obliged to clarify it, according to two employees who heard the remarks, who asked to remain anonymous because they're not authorized to speak to the press.

Pichai defined "Googleyness" as the following, per one of those sources:

"Mission First"

"Make Helpful Things"

"Be Bold & Responsible"

"Stay Scrappy"

"Hustle & Have Fun"

"Team Google"

A Google spokesperson declined to comment.

The term "Googleyness" has always been amorphous. In his 2015 book Work Rules, Google's former head of people operations, Laszlo Block, listed certain attributes that he considered "Googleyness," such as "intellectual humility," "enjoying fun," and "comfort with ambiguity."

The company previously changed its hiring guidelines to "avoid confusing Googleyness with culture fit," The Information reported in 2019. The change came after the company had been criticized for its lack of diversity in its workplace.

Are you a current or former Google employee with something to share? You can reach the reporter Hugh Langley via the encrypted messaging app Signal (+1-628-228-1836) or email ([email protected]).

• Suchir Balaji helped OpenAI collect data from the internet for AI model training, the NYT reported.
• He was found dead in an apartment in San Francisco in late November, according to police.
• About a month before, Balaji published an essay criticizing how AI models use data.

The recent death of former OpenAI researcher Suchir Balaji has brought an under-discussed AI debate back into the limelight.

AI models are trained on information from the internet. These tools answer user questions directly, so fewer people visit the websites that created and verified the original data. This drains resources from content creators, which could lead to a less accurate and rich internet.

Elon Musk calls this "Death by LLM." Stack Overflow, a coding Q&A website, has already been damaged by this phenomenon. And Balaji was concerned about this.

Balaji was found dead in late November. The San Francisco Police Department said it found "no evidence of foul play" during the initial investigation. The city's chief medical examiner determined the death to be suicide.

Balaji's concerns

About a month before Balaji died, he published an essay on his personal website that addressed how AI models are created and how this may be bad for the internet. 

He cited research that studied the impact of AI models using online data for free to answer questions directly while sucking traffic away from the original sources.

The study analyzed Stack Overflow and found that traffic to this site declined by about 12% after the release of ChatGPT. Instead of going to Stack Overflow to ask coding questions and do research, some developers were just asking ChatGPT for the answers. 

Other findings from the research Balaji cited: 

• There was a decline in the number of questions posted on Stack Overflow after the release of ChatGPT.
• The average account age of the question-askers rose after ChatGPT came out, suggesting fewer people signed up to Stack Overflow or that more users left the online community.

This suggests that AI models could undermine some of the incentives that created the information-rich internet as we know it today.

If people can get their answers directly from AI models, there's no need to go to the original sources of the information. If people don't visit websites as much, advertising and subscription revenue may fall, and there would be less money to fund the creation and verification of high-quality online data.

MKBHD wants to opt out

It's even more galling to imagine that AI models might be doing this based partly on your own work. 

Tech reviewer Marques Brownlee experienced this recently when he reviewed OpenAI's Sora video model and found that it created a clip with a plant that looked a lot like a plant from his own videos posted on YouTube. 

"Are my videos in that source material? Is this exact plant part of the source material? Is it just a coincidence?" said Brownlee, who's known as MKBHD.

Naturally, he also wanted to know if he could opt out and prevent his videos from being used to train AI models. "We don't know if it's too late to opt out," Brownlee said.

'Not a sustainable model'

In an interview with The New York Times published in October, Balaji said AI chatbots like ChatGPT are stripping away the commercial value of people's work and services.

The publication reported that while working at OpenAI, Balaji was part of a team that collected data from the internet for AI model training. He joined the startup with high hopes for how AI could help society, but became disillusioned, NYT wrote. 

"This is not a sustainable model for the internet ecosystem," he told the publication.

In a statement to the Times about Balaji's comments, OpenAI said the way it builds AI models is protected by fair use copyright principles and supported by legal precedents. "We view this principle as fair to creators, necessary for innovators, and critical for US competitiveness," it added.

In his essay, Balaji disagreed.

One of the four tests for copyright infringement is whether a new work impacts the potential market for, or value of, the original copyrighted work. If it does this type of damage, then it's not "fair use" and is not allowed. 

Balaji concluded that ChatGPT and other AI models don't quality for fair use copyright protection. 

"None of the four factors seem to weigh in favor of ChatGPT being a fair use of its training data," he wrote. "That being said, none of the arguments here are fundamentally specific to ChatGPT either, and similar arguments could be made for many generative AI products in a wide variety of domains."

Talking about data

Tech companies producing these powerful AI models don't like to talk about the value of training data. They've even stopped disclosing where they get the data from, which was a common practice until a few years ago. 

"They always highlight their clever algorithms, not the underlying data," Nick Vincent, an AI researcher, told BI last year.

Balaji's death may finally give this debate the attention it deserves. 

"We are devastated to learn of this incredibly sad news today and our hearts go out to Suchir's loved ones during this difficult time," an OpenAI spokesperson told BI recently. 

If you or someone you know is experiencing depression or has had thoughts of harming themself or taking their own life, get help. In the US, call or text 988 to reach the Suicide & Crisis Lifeline, which provides 24/7, free, confidential support for people in distress, as well as best practices for professionals and resources to aid in prevention and crisis situations. Help is also available through the Crisis Text Line — just text "HOME" to 741741. The International Association for Suicide Prevention offers resources for those outside the US.

It's been a really busy month for Google as it apparently endeavors to outshine OpenAI with a blitz of AI releases. On Thursday, Google dropped its latest party trick: Gemini 2.0 Flash Thinking Experimental, which is a new AI model that uses runtime "reasoning" techniques similar to OpenAI's o1 to achieve "deeper thinking" on problems fed into it.

The experimental model builds on Google's newly released Gemini 2.0 Flash and runs on its AI Studio platform, but early tests conducted by TechCrunch reporter Kyle Wiggers reveal accuracy issues with some basic tasks, such as incorrectly counting that the word "strawberry" contains two R's.

These so-called reasoning models differ from standard AI models by incorporating feedback loops of self-checking mechanisms, similar to techniques we first saw in early 2023 with hobbyist projects like "Baby AGI." The process requires more computing time, often adding extra seconds or minutes to response times. Companies have turned to reasoning models as traditional scaling methods at training time have been showing diminishing returns.

Read full article

Comments

For months, popular fighting game YouTubers have been under attack. Even the seemingly most cautious among them have been duped by sophisticated phishing attacks that hack their accounts to push cryptocurrency scams by convincingly appearing to offer legitimate sponsorships from established brands.

These scams often start with bad actors seemingly taking over verified accounts on X (formerly Twitter) with substantial followings and then using them to impersonate marketing managers at real brands who can be easily found on LinkedIn.

The fake X accounts go to great lengths to appear legitimate. They link to brands' actual websites and populate feeds with histories seemingly spanning decades by re-posting brands' authentic posts.

Read full article

Comments

Google has released what it’s calling a new “reasoning” AI model — but it’s in the experimental stages, and from our brief testing, there’s certainly room for improvement. The new model, called Gemini 2.0 Flash Thinking Experimental (a mouthful, to be sure), is available in AI Studio, Google’s AI prototyping platform. A model card describes […]

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

The AI-generated video scene has been hopping this year (or twirling wildly, as the case may be). This past week alone we've seen releases or announcements of OpenAI's Sora, Pika AI's Pika 2, Google's Veo 2, and Minimax's video-01-live. It's frankly hard to keep up, and even tougher to test them all. But recently, we put a new open-weights AI video synthesis model, Tencent's HunyuanVideo, to the test—and it's surprisingly capable for being a "free" model.

Unlike the aforementioned models, HunyuanVideo's neural network weights are openly distributed, which means they can be run locally under the right circumstances (people have already demonstrated it on a consumer 24 GB VRAM GPU) and it can be fine-tuned or used with LoRAs to teach it new concepts.

Notably, a few Chinese companies have been at the forefront of AI video for most of this year, and some experts speculate that the reason is less reticence to train on copyrighted materials, use images and names of famous celebrities, and incorporate some uncensored video sources. As we saw with Stable Diffusion 3's mangled release, including nudity or pornography in training data may allow these models achieve better results by providing more information about human bodies. HunyuanVideo notably allows uncensored outputs, so unlike the commercial video models out there, it can generate videos of anatomically realistic, nude humans.

Read full article

Comments

You might be tempted to get a Chromebook as your next laptop and we wouldn’t blame you. Chrome OS has come a long way: Chromebooks today have a good combination of build quality, battery life and power while also supporting years of software updates that will take you years into the future. If you primarily browse the web, or don’t need a ton of extra software on a regular basis, a Chromebook can be a powerful daily driver.

While Google did make things simpler last fall by introducing the Chromebook Plus initiative (more on that below), there are still multiple things to keep in mind when shopping for a new Chromebook. The market is saturated with machines at different price points and power levels; I’ve been testing and reviewing Chromebooks for years and know what you should be looking for, and what you can expect out of these laptops.

What is Chrome OS, and why would I use it over Windows?

This is probably the number one question about Chromebooks. There are plenty of inexpensive Windows laptops on the market, so why bother with Chrome's operating system? Glad you asked. For me, the simple and clean nature of Chrome OS is a big selling point. Chrome OS is based on Google’s Chrome browser, which means most of the programs you can run are web based. There’s no bloatware or unwanted apps to uninstall like you often get on Windows laptops, it boots up in seconds, and you can completely reset to factory settings almost as quickly.

Of course, simplicity will also be a major drawback for some users. Not being able to install native software can be a dealbreaker if you’re a video editor or software developer. But there are also plenty of people who do the majority of their work in a web browser.

Google and its software partners are getting better every year at supporting more advanced features. For example, Google added video editing tools to the Google Photos app on Chromebooks – it won’t replace Adobe Premiere, but it should be handy for a lot of people. Similarly, Google and Adobe announced Photoshop on the web last year, something that brings much of the power of Adobe’s desktop apps to Chromebooks.

Chromebooks can also run Android apps, which greatly expands the amount of software available. The quality varies widely, but it means you can do more with a Chromebook beyond just web-based apps. For example, you can install the Netflix app and save videos for offline watching. Other Android apps like Microsoft Office and Adobe Lightroom are surprisingly capable as well. Between Android apps and a general improvement in web apps, Chromebooks are more than just portals to a browser.

What do Chromebooks do well?

Photo by Nathan Ingraham / Engadget

Put simply, web browsing and really anything web based. Online shopping, streaming music and video and using various social media sites are among the most common daily tasks people do on Chromebooks. As you might expect, they also work well with Google services like Photos, Docs, Gmail, Drive, Keep and so on. Yes, any computer that can run Chrome can do that too, but the lightweight nature of Google Chrome OS makes it a responsive and stable platform.

As I mentioned before, Chrome OS can run Android apps, so if you’re an Android user you’ll find some nice ties between the platforms. You can get most of the same apps that are on your phone on a Chromebook and keep info in sync between them. You can also use some Android phones as a security key for your Chromebook or instantly tether your laptop to use mobile data.

Google continues to tout security as a major differentiator for Chromebooks, and it’s definitely a factor worth considering. Auto-updates are the first lines of defense: Chrome OS updates download quickly in the background and a fast reboot is all it takes to install the latest version. Google says that each webpage and app on a Chromebook runs in its own sandbox as well, so any security threats are contained to that individual app. Finally, Chrome OS has a self-check called Verified Boot that runs every time a device starts up. Beyond all this, the simple fact that you generally can’t install traditional apps on a Chromebook means there are fewer ways for bad actors to access the system.

If you’re interested in Google’s Gemini AI tools, a Chromebook is a good option as well. Every Chromebook in our top picks comes with a full year of the Google One AI Premium plan — this combines the usual Google One perks like 2TB of storage and 10 percent back in purchases from the Google Store with a bunch of AI tools. You’ll get access to Gemini in Gmail, Docs and other apps, Gemini Advanced (which runs on the 1.5 Pro model) and more. Given that this plan is $20/month, it’s a pretty solid perk. Chromebook Plus models also include tools like the AI-powered “help me write,” the Google Photos Magic Editor and generative AI backgrounds you can create by filling in a few prompts.

As for when to avoid Chromebooks, the answer is simple: If you rely heavily on a specific native application for Windows or a Mac, chances are you won’t find the exact same option on a ChromeOS device. That’s most true in fields like photo and video editing, but it can also be the case in law or finance. Plenty of businesses run on Google’s G suite software, but more still have specific requirements that a Chromebook might not match. If you’re an iPhone user, you’ll also miss out on the way the iPhone easily integrates with an iPad or Mac. For me, the big downside is not being able to access iMessage on a Chromebook.

Finally, gaming Chromebooks are not ubiquitous, although they’re becoming a slightly more reasonable option with the rise of cloud gaming. In late 2022, Google and some hardware partners announced a push to make Chromebooks with cloud gaming in mind. From a hardware perspective, that means laptops with bigger screens that have higher refresh rates as well as optimizing those laptops to work with services like NVIDIA GeForce Now, Xbox Game Pass and Amazon Luna. You’ll obviously need an internet connection to use these services, but the good news is that playing modern games on a Chromebook isn’t impossible. You can also install Android games from the Google Play Store, but that’s not what most people are thinking of when they want to game on a laptop.

What are the most important specs for a Chromebook?

Photo by Nathan Ingraham / Engadget

Chrome OS is lightweight and runs well on fairly modest hardware, so the most important thing to look for might not be processor power or storage space. But Google made it easier to get consistent specs and performance late last year when it introduced the Chromebook Plus initiative. Any device with a Chromebook Plus designation meets some minimum requirements, which happen to be very similar to what I’d recommend most people get if they’re looking for a laptop they can use every day.

Chromebook Plus models have at least a 12th-gen Intel Core i3 processor, or an AMD Ryzen 3 7000 series processor, both of which should be more than enough for most people. These laptops also have a minimum of 8GB of RAM and 128GB of storage, which should do the trick unless you’re really pushing your Chromebook. All Chromebook Plus models have to have a 1080p webcam, which is nice in these days of constant video calling, and they also all have to have at least a 1080p IPS screen.

Of course, you can get higher specs or better screens if you desire, but I’ve found that basically everything included in the Chromebook Plus target specs makes for a very good experience.

Google has an Auto Update policy for Chromebooks as well, and while that’s not exactly a spec, it’s worth checking before you buy. Last year, Google announced that Chromebooks would get software updates and support for an impressive 10 years after their release date. This support page lists the Auto Update expiration date for virtually every Chromebook ever, but a good rule of thumb is to buy the newest machine you can to maximize your support.

How much should I spend?

Chromebooks started out notoriously cheap, with list prices often coming in under $300. But as they’ve gone more mainstream, they’ve transitioned from being essentially modern netbooks to the kind of laptop you’ll want to use all day. As such, prices have increased: At this point, you should expect to spend at least $400 if you want a solid daily driver. There are still many cheap Chromebooks out there that may be suitable as secondary devices, but a good Chromebook that can be an all-day, every-day laptop will cost more. But, notably, even the best Chromebooks usually cost less than the best Windows laptops, or even the best “regular” laptops out there.

There are plenty of premium Chromebooks that approach or even exceed $1,000 that claim to offer better performance and more processing power, but I don’t recommend spending that much. Generally, that’ll get you a better design with more premium materials, as well as more powerful internals and extra storage space. Of course, you also sometimes pay for the brand name. But, the specs I outlined earlier are usually enough.

See Also:

• Best Laptops for 2025

• Best Gaming Laptops

• Best 2-in-1 Laptops for 2025

• Best Cheap Windows Laptops

• Best Laptops for College Students

Best Chromebooks in 2025

Other Chromebooks we tested

Samsung Galaxy Chromebook Plus

Samsung’s new Galaxy Chromebook Plus is one of the more unique Chromebooks out there. It’s extremely thin and light, at .46 inches and 2.6 pounds, but it manages to include a 15.6-inch display in that frame. That screen is a 1080p panel that’s sharp and bright, but its 16:9 aspect ratio made things feel a bit cramped when scrolling vertically. Performance is very good, and the keyboard is solid, though I’m not a fan of the number pad as it shifts everything to the left. At $700 it’s not cheap, but that feels fair considering its size and capabilities. If you’re looking for a big screen laptop that is also super light, this Chromebook merits consideration, even if it’s not the best option for everyone.

• Staff at major companies have asked their leaders if there are plans to follow Amazon's full return to office.
• Firms like Meta, Google, and Microsoft have a hybrid setup — however, execs say they're eyeing productivity.
• Research findings on the subject are varied, and the debate will likely continue in 2025.

Executives at major companies are referencing a specific term to hedge when asked by employees if they plan to follow in Amazon's footsteps and implement a return to 5 days a week in the office.

That word? Productivity.

While Amazon has been the most high-profile example this year of a full return to office policy, set to go into effect in January, telecom giant AT&T has also elected to double down on in-person work with a similar 5-day policy, Business Insider first reported.

In the wake of Amazon's announcement, executives at both Google and Microsoft, which require employees to be in the office at least 3 days a week, have fielded questions from staff wondering if the days of hybrid work are numbered.

Microsoft's executive vice president of cloud and AI, Scott Guthrie, said the company wouldn't change the hybrid work policy unless it noticed a drop in productivity, BI reported in September.

In October, Google CEO Sundar Pichai said the company had no plans to order employees back to the office, so long as employees remain productive during their at-home work days, BI previously reported.

Over at Meta, Mark Zuckerberg said last year that "early analysis of performance data," indicated productivity increases for early-career engineers in the office at least 3 days a week. A few months later, the company announced it was requiring employees to return to the office 3 days a week.

Executives at Dell called the company's sales team back to the office 5 days a week starting at the end of September, writing in a memo, "Our data shows that sales teams are more productive when onsite."

Though Amazon did not explicitly name productivity as a reason for its full return to the office, CEO Andy Jassy emphasized a similar term: effectiveness.

Being back in person 5 days a week makes it "easier for our teammates to learn, model, practice, and strengthen our culture; collaborating, brainstorming, and inventing are simpler and more effective; teaching and learning from one another are more seamless; and, teams tend to be better connected to one another," he wrote at the time.

For those committing to a full return to office, preparing campuses for the influx of employees in the new year is its own challenge. Amazon has since delayed the announced January 2 effective date of the new mandate for some employees because it doesn't have enough office space in some locations, BI reported earlier this month.

As CEOs and company leaders keep an eye on how employees in remote or hybrid setups perform, various studies since the onset of the pandemic have attempted to measure and compare the productivity of employees who work at home and in-office. Research studies have produced conflicting results, further complicated by the matter of how best to define or measure productivity.

Goldman Sachs, which has a 5-day-in-office policy, reviewed several analyses that used different ways of evaluating changes in work-from-home productivity, from call-center workers who were randomly chosen to work from home to comparing the productivity of randomly assigned remote workers with their in-office peers.

In short, it's hard to say for sure, and executives are deciding what their long-term setup will be after a year in which some of the world's biggest companies put a renewed focus on being "lean" and "efficient."

Meanwhile, some employees have returned to commuting in (sometimes "coffee-badging" in and returning home), others have relocated to comply with a policy change, and some have resigned to pursue a hybrid or fully remote opportunity. As companies tighten their belts and conduct layoffs, other workers have taken to workplace forums to wonder if some of the RTO mandates have been a possible "quiet layoffs" tactic.

As more major global companies revisit their policies and make changes, CEOs are likely to face more questions on the topic going into the new year.

For some, the answer is simple: Stay productive and we'll stay flexible.

• Motorola Solutions uses AI to help address delays in 911 emergency calls and improve response times.
• Its Vesta NXT software helps 911 call handlers gather and summarize data for quicker communication.
• This article is part of "CXO AI Playbook" — straight talk from business leaders on how they're testing and using AI.

Motorola Solutions is a Chicago-based provider of technology and communications solutions focused on public safety and enterprise security. It has about 21,000 employees worldwide.

Situation analysis: What problem was the company trying to solve?

The National Emergency Number Association estimates that 240 million 911 calls are made in the US each year. But fragmented emergency-response systems across various agencies and organizations can lead to dangerous delays.

"You hope to never call 911, but when you do, it needs to work," Jehan Wickramasuriya, the corporate vice president of AI and platforms at Motorola Solutions, told Business Insider.

He added that call takers' jobs can be demanding and unpredictable, and they're often under intense pressure. "There can be a high level of stress if there's an active shooter or domestic disturbance," he said. "They're trying to keep a caller calm and simultaneously find out if they need medical help." Meanwhile, he said, callers may be "speaking so fast that it's difficult to understand and retain everything they say."

Pinpointing a caller's location adds a layer of complexity. Mobile 911 calls are typically routed based on cell-tower locations rather than the caller's actual position. This requires calls to be redirected, adding several seconds to response times.

"At the end of the day it's a data problem," Wickramasuriya said, "because a lot of information needs to get transmitted in each call."

Motorola Solutions is using AI to consolidate this data in a single platform.

Key staff and stakeholders

The company structures its AI research team around specialized AI domains, such as computer vision and speech and audio processing, rather than individual product lines.

Wickramasuriya said the core AI team consisted of about 50 scientists, developers, and engineers who collaborate closely with hundreds of product managers, designers, and user-experience specialists.

Motorola Solutions also works with various cloud and technology vendors on its AI-enabled products and services.

AI in action

In June, Motorola Solutions launched Vesta NXT, software designed to help 911 call handlers manage emergency calls. It brings data from various public-safety systems onto one platform, helping the handlers gather and summarize information.

The tool uses AI to surface details including the caller's location and, for callers who have opted to share their medical profile from their phone, any underlying health conditions. It can also suggest the best entrance to a building. "That's important information for first responders," Wickramasuriya said.

The software has translation and transcription capabilities, helping English speakers and non-English speakers communicate. AI also helps call handlers manage nonemergency calls — by streamlining the reporting of issues like abandoned cars or stolen property, call handlers can focus more on critical emergencies.

Most important, AI can improve the human element of emergency response. "AI is working in the background to help the call taker attend to the person on the other end of the line," Wickramasuriya said.

Did it work, and how did leaders know?

Motorola Solutions says roughly 60% of 911 call centers in the US use its call-handling software. It's transitioning existing Vesta 911 users to its new system with the AI features.

The company says these AI tools are already translating millions of minutes of audio each month and have helped lighten emergency-call handlers' workloads partly by resolving nonemergency calls and connecting callers to other resources.

Lee County is the first Public Safety Answering Point, which is a call center that handles emergency calls and coordinates responses, to use the VESTA NXT. Motorola Solutions said administrators there found the AI-generated searchable text transcripts and real-time summaries of 911 calls that call handlers can share with dispatchers and first responders helped save time and alleviate stress for call handlers.

What's next?

Wickramasuriya said the company was focused on improving Vesta NXT.

He said the goal was to "expand the usefulness" of the software by integrating it more deeply into existing workflows, including by developing features that connect first responders directly with dispatchers and call takers.

Another aim, he said, is to help understaffed 911 call centers "understand their staffing needs and identify which call takers are handling high-stress situations and address stress and fatigue among call handlers."

• Early testers of Google's new AI video generator compared it with OpenAI's Sora.
• So far, Google's results are blowing OpenAI's out of the water.
• Google has tapped YouTube to train its AI models but says other companies can't do the same.

Not to let OpenAI have all the fun with its 12 days of 'Shipmas,' Google on Monday revealed its new AI video generator, Veo 2. Early testers found it's blowing OpenAI's Sora out of the water.

OpenAI made its Sora AI video generator available for general use earlier this month, while Google's is still in early preview. Still, people are sharing comparisons of the two models running the same prompts, and so far, Veo has proved more impressive.

Why is Google's Veo 2 doing better than Sora so far? The answer may be YouTube, which Google owns and has used to train these models.

TED host and former Googler Bilawal Sidhu shared some comparisons of Google's Veo 2 and OpenAI's Sora on X.

He said he used the prompt, "Eating soup like they do in Europe, the old fashioned way," which generated a terrifying result in Sora and something more impressive in Google's Veo 2.

Here's another, which took a prompt that YouTube star Marques Brownlee had tried in a video reviewing Sora.

This one from EasyGen founder Ruben Hassid included a prompt for someone cutting a tomato with a knife. In the video he shared, we see how Google's Veo 2 had the knife going cleanly through the tomato and avoiding fingers, while the knife in Sora's video cut through the hand.

Granted, these are cherry-picked examples, but the consensus among AI enthusiasts is that Google has outperformed.

Andreessen Horowitz partner Justine Moore wrote on X that she had spent a few hours testing Veo against Sora and believed Sora's "biases towards more motion" while Veo focuses on accuracy and physics.

Google has been open about tapping YouTube data for its AI, but it does not permit others to do the same. The New York Times previously reported that OpenAI had trained its models using some YouTube data anyway. YouTube CEO Neal Mohan said OpenAI doing this would violate Google's policies.

BI previously reported that Google's DeepMind also tapped YouTube's vast trove of content to build an impressive music generator that never saw the light of day.

Google did not immediately respond to a request for comment.

Are you a Google or OpenAI employee? Got more insight to share? You can reach the reporter Hugh Langley via the encrypted messaging app Signal (+1-628-228-1836) or email ([email protected]).