Technologies

AI Gets Smarter, Safer, More Visual With GPT-4 Update, OpenAI Says

If you subscribe to ChatGPT Plus, you can try it out now.

The hottest AI technology foundation got a big upgrade Tuesday with OpenAI’s GPT-4 release now available in the premium version of the ChatGPT chatbot.

GPT-4 can generate much longer strings of text and respond when people feed it images, and it’s designed to do a better job avoiding artificial intelligence pitfalls visible in the earlier GPT-3.5, OpenAI said Tuesday. For example, when taking bar exams that attorneys must pass to practice law, GPT-4 ranks in the top 10% of scores compared with the bottom 10% for GPT-3.5, the AI research company said.

GPT stands for Generative Pretrained Transformer, a reference to the fact that it can generate text on its own — now up to 25,000 words with GPT-4 — and that it uses an AI technology called transformers that Google pioneered. It’s a type of AI called a large language model, or LLM, that’s trained on vast swaths of data harvested from the internet, learning mathematically to spot patterns and reproduce styles. Human overseers rate results to steer GPT in the right direction, and GPT-4 has more of this feedback.

OpenAI has made GPT available to developers for years, but ChatGPT, which debuted in November, offered an easy interface ordinary folks can use. That yielded an explosion of interest, experimentation and worry about the downsides of the technology. It can do everything from generating programming code and answering exam questions to writing poetry and supplying basic facts. It’s remarkable if not always reliable.

ChatGPT is free, but it can falter when demand is high. In January, OpenAI began offering ChatGPT Plus for $20 per month with assured availability and, now, the GPT-4 foundation. Developers can sign up on a waiting list to get their own access to GPT-4.

GPT-4 advancements

«In a casual conversation, the distinction between GPT-3.5 and GPT-4 can be subtle. The difference comes out when the complexity of the task reaches a sufficient threshold,» OpenAI said. «GPT-4 is more reliable, creative and able to handle much more nuanced instructions than GPT-3.5.»

Another major advance in GPT-4 is the ability to accept input data that includes text and photos. OpenAI’s example is asking the chatbot to explain a joke showing a bulky decades-old computer cable plugged into a modern iPhone’s tiny Lightning port. This feature also helps GPT take tests that aren’t just textual, but it isn’t yet available in ChatGPT Plus.

Another is better performance avoiding AI problems like hallucinations — incorrectly fabricated responses, often offered with just as much seeming authority as answers the AI gets right. GPT-4 also is better at thwarting attempts to get it to say the wrong thing: «GPT-4 scores 40% higher than our latest GPT-3.5 on our internal adversarial factuality evaluations,» OpenAI said.

GPT-4 also adds new «steerability» options. Users of large language models today often must engage in elaborate «prompt engineering,» learning how to embed specific cues in their prompts to get the right sort of responses. GPT-4 adds a system command option that lets users set a specific tone or style, for example programming code or a Socratic tutor: «You are a tutor that always responds in the Socratic style. You never give the student the answer, but always try to ask just the right question to help them learn to think for themselves.»

«Stochastic parrots» and other problems

OpenAI acknowledges significant shortcomings that persist with GPT-4, though it also touts progress avoiding them.

«It can sometimes make simple reasoning errors … or be overly gullible in accepting obvious false statements from a user. And sometimes it can fail at hard problems the same way humans do, such as introducing security vulnerabilities into code it produces,» OpenAI said. In addition, «GPT-4 can also be confidently wrong in its predictions, not taking care to double-check work when it’s likely to make a mistake.»

Large language models can deliver impressive results, seeming to understand huge amounts of subject matter and to converse in human-sounding if somewhat stilted language. Fundamentally, though, LLM AIs don’t really know anything. They’re just able to string words together in statistically very refined ways.

This statistical but fundamentally somewhat hollow approach to knowledge led researchers, including former Google AI researchers Emily Bender and Timnit Gebru, to warn of the «dangers of stochastic parrots» that come with large language models. Language model AIs tend to encode biases, stereotypes and negative sentiment present in training data, and researchers and other people using these models tend «to mistake … performance gains for actual natural language understanding.»

OpenAI Chief Executive Sam Altman acknowledges problems, but he’s pleased overall with the progress shown with GPT-4. «It is more creative than previous models, it hallucinates significantly less, and it is less biased. It can pass a bar exam and score a 5 on several AP exams,» Altman tweeted Tuesday.

One worry about AI is that students will use it to cheat, for example when answering essay questions. It’s a real risk, though some educators actively embrace LLMs as a tool, like search engines and Wikipedia. Plagiarism detection companies are adapting to AI by training their own detection models. One such company, Crossplag, said Wednesday that after testing about 50 documents that GPT-4 generated, «our accuracy rate was above 98.5%.»

OpenAI, Microsoft and Nvidia partnership

OpenAI got a big boost when Microsoft said in February it’s using GPT technology in its Bing search engine, including a chat features similar to ChatGPT. On Tuesday, Microsoft said it’s using GPT-4 for the Bing work. Together, OpenAI and Microsoft pose a major search threat to Google, but Google has its own large language model technology too, including a chatbot called Bard that Google is testing privately.

Also on Tuesday, Google announced it’ll begin limited testing of its own AI technology to boost writing Gmail emails and Google Docs word processing documents. «With your collaborative AI partner you can continue to refine and edit, getting more suggestions as needed,» Google said.

That phrasing mirrors Microsoft’s «co-pilot» positioning of AI technology. Calling it an aid to human-led work is a common stance, given the problems of the technology and the necessity for careful human oversight.

Microsoft uses GPT technology both to evaluate the searches people type into Bing and, in some cases, to offer more elaborate, conversational responses. The results can be much more informative than those of earlier search engines, but the more conversational interface that can be invoked as an option has had problems that make it look unhinged.

To train GPT, OpenAI used Microsoft’s Azure cloud computing service, including thousands of Nvidia’s A100 graphics processing units, or GPUs, yoked together. Azure now can use Nvidia’s new H100 processors, which include specific circuitry to accelerate AI transformer calculations.

AI chatbots everywhere

Another large language model developer, Anthropic, also unveiled an AI chatbot called Claude on Tuesday. The company, which counts Google as an investor, opened a waiting list for Claude.

«Claude is capable of a wide variety of conversational and text processing tasks while maintaining a high degree of reliability and predictability,» Anthropic said in a blog post. «Claude can help with use cases including summarization, search, creative and collaborative writing, Q&A, coding and more.»

It’s one of a growing crowd. Chinese search and tech giant Baidu is working on a chatbot called Ernie Bot. Meta, parent of Facebook and Instagram, consolidated its AI operations into a bigger team and plans to build more generative AI into its products. Even Snapchat is getting in on the game with a GPT-based chatbot called My AI.

Expect more refinements in the future.

«We have had the initial training of GPT-4 done for quite awhile, but it’s taken us a long time and a lot of work to feel ready to release it,» Altman tweeted. «We hope you enjoy it and we really appreciate feedback on its shortcomings.»

Editors’ note: CNET is using an AI engine to create some personalfinance explainers that are edited and fact-checked by our editors. Formore, see this post.

Technologies

T-Mobile Adds New Top 5G Plans, T-Satellite and New 5-Year Price Locks

The new top unlimited plans, Experience More and Experience Beyond, shave some costs and add data and satellite options.

Just two years after expanding its lineup of cellular plans, T-Mobile this week announced two new plans that replace its Go5G Plus and Go5G Next offerings, refreshed its prepaid Metro line and wrapped them all in a promised five-year pricing guarantee.

To convert more subscribers, the carrier is also offering up to $800 to help customers pay off phone balances when switching from another carrier.

In a briefing with CNET, Jon Friar, president of T-Mobile’s consumer group, explained why the company is revamping and simplifying its array of mobile plans. «The pain point that’s out there over the last couple of years is rising costs all around consumers,» Friar said. «For us to be able to bring more value and even lower prices on [plans like] Experience More versus our former Go5G Plus is a huge win for consumers.»

The new plans went into effect April 23.

With these changes, CNET is already hard at work updating our picks for Best T-Mobile Plans, so check back soon for our recommendations.

More Experiences to define the T-Mobile experience

The top of the new T-Mobile postpaid lineup is two new plans: Experience More and Experience Beyond.

Experience More is the next generation of the Go5G Plus plan, which has unlimited 5G and 4G LTE access and unlimited Premium Data (download speeds up to 418Mbps and upload speeds up to 31Mbps). High-speed hotspot data is bumped up to 60GB from 50GB per month. The monthly price is now $5 lower per line than Go5G Plus.

The Experience More plan also gets free T-Satellite with Starlink service (the new name for T-Mobile’s satellite feature that uses Starlink’s constellation of satellites) through the end of 2025. Although T-Satellite is still officially in beta until July, customers can continue to get free access to the beta starting now. At the start of the new year, the service will cost $10 per month, a $5 drop from T-Mobile’s originally announced pricing. T-Satellite will be open to customers of other carriers for the same pricing beginning in July.

The new top-tier plan, Experience Beyond, also comes in $5 per line cheaper than its predecessor, Go5G Next. It has 250GB of high-speed hotspot data per month, up from 50GB, and more data when you’re traveling outside the US: 30GB in Canada and Mexico (versus 15GB) and 15GB in 215 countries (up from 5GB). T-Satellite service is included in the Experience Beyond plan.

However, one small change to the Experience plans affects that pricing: Taxes and fees, previously included in the Go5G Plus and Go5G Next prices, are now broken out separately. T-Mobile recently announced that one such fee, the Regulatory Programs and Telco Recovery Fee, would increase up to 50 cents per month.

According to T-Mobile, the Experience Beyond rates and features will be «rolling out soon» for customers currently on the Go5G Next plan.

The Essentials plan is staying in the lineup at the same cost of $60 per month for a single line, the same 50GB of Premium Data and unlimited 5G and 4G LTE data. High-speed hotspot data is an optional $10 add-on, as is T-Satellite access, for $15 (both per month).

Also still in the mix is the Essentials Saver plan, an affordable option that has ranked high in CNET’s Best Cellphone Plans recommendations.

Corresponding T-Mobile plans, such as those for military, first responders and people age 55 and older are also getting refreshed with the new lineup.

T-Mobile’s plan shakeup is being driven in part by the current economic climate. Explaining the rationale behind the price reductions and the streamlined number of plans, Mike Katz, president of marketing, innovation and experience at T-Mobile told CNET, «We’re in a weird time right now where prices everywhere are going up and they’ve happened over the last several years. We felt like there was an opportunity to compete with some simplicity, but more importantly, some peace of mind for customers.»

Existing customers who want to switch to one of the new plans can do so at the same rates offered to new customers. Or, if a current plan still works for them, they can continue without changes (although keep in mind that T-Mobile earlier this year increased prices for some legacy plans).

Five years of price stability

It’s nearly impossible to think about prices these days without warily eyeing how tariffs and US economic policy will affect what we pay for things. So it’s not surprising to see carriers implement some cost stability into their plans. For instance, Verizon recently locked prices for three years on their plans.

Now, T-Mobile is building a five-year price guarantee for its T-Mobile and Metro plans. That pricing applies to talk, text and data amounts — not necessarily taxes and other fees that can fluctuate.

Given the uncertain outlook, it seems counterintuitive to lock in a longer rate. When asked about this, Katz said, «We feel like our job is to solve pain points for customers and we feel like this helps with this exact sentiment. It shifts the risk from customers to us. We’ll take the risk so they don’t have to.»

The price hold applies to new customers signing up for the plans as well as current customers switching to one. T-Mobile is offering the same deals and pricing to new and existing subscribers. Also, the five-year deal applies to pricing; it’s not a five-year plan commitment.

More money and options to encourage switchers

The promise of a five-year price guarantee is also intended to lure people from other carriers, particularly AT&T and Verizon. As further incentive, T-Mobile is offering up to $800 per line (distributed via a virtual prepaid Mastercard) to help pay off other carriers’ device contracts. This is a limited-time offer. There are also options to trade in old devices, including locked phones, to get up to four new flagship phones.

Or, if getting out of a contract isn’t an issue, T-Mobile can offer $200 in credit (up to $800 for four lines) to bring an existing number to the network.

Four new Metro prepaid plans

On the prepaid side, T-Mobile is rolling out four new Metro plans, which are also covered by the new five-year price guarantee:

• Metro Starter costs $25 per line per month for a family of four and there is no need to bring an existing number. (The cost is $105 the first month.)

• Metro Starter Plus runs $40 per month for a new phone, unlimited talk, text and 5G data when bringing an existing number. For $65 per month, new customers can get two lines and two new Samsung A15 phones. No autopay is required.

• Metro Flex Unlimited is $30 per line per month with autopay for four lines ($125 the first month) with unlimited talk, text and 5G data.

• Metro Flex Unlimited Plus costs $60 per line per month, then $35 for lines two and three and then lowers the price of the fourth line to $10 per month as more family members are added. Adding a tablet or smartwatch to an existing line costs $5. And streaming video, such as from the included Amazon Prime membership, comes through at HD quality.

See more: If you’re looking for phone plans, you may also be looking for a new cell phone. Here are CNET’s picks.

The Pixel 9A’s Design: Google Takes Minimalism to the Extreme

+7 More

See all photos

Technologies

Did Apple Just Save the iPhone From Extreme Tariff Price Hikes?

Technologies

Today’s NYT Mini Crossword Answers for Saturday, April 26

Here are the answers for The New York Times Mini Crossword for April 26.

Looking for the most recent Mini Crossword answer? Click here for today’s Mini Crossword hints, as well as our daily answers and hints for The New York Times Wordle, Strands, Connections and Connections: Sports Edition puzzles.

Today’s NYT Mini Crossword could be tricky. I, for one, did NOT know the country in which «the automobile, aspirin and the accordion were invented,» but as the other clues filled in, I figured it out. Need some help with today’s Mini Crossword? Read on. And if you could use some hints and guidance for daily solving, check out our Mini Crossword tips.

The Mini Crossword is just one of many games in the Times’ games collection. If you’re looking for today’s Wordle, Connections, Connections: Sports Edition and Strands answers, you can visit CNET’s NYT puzzle hints page.

Read more: Tips and Tricks for Solving The New York Times Mini Crossword

Let’s get at those Mini Crossword clues and answers.

Mini across clues and answers

1A clue: Fancy party
Answer: GALA

5A clue: Five ___ (discount store chain)
Answer: BELOW

6A clue: Bottom stripe on a pride flag
Answer: PURPLE

7A clue: Bedtime outfit, informally
Answer: JAMMIES

8A clue: Task to «run»
Answer: ERRAND

9A clue: Actress Keaton of «Father of the Bride»
Answer: DIANE

10A clue: Car ride game
Answer: ISPY

Mini down clues and answers

1D clue: Country in which the automobile, aspirin and the accordion were invented
Answer: GERMANY

2D clue: Mountain-related
Answer: ALPINE

3D clue: Was audibly amused, slangily
Answer: LOLED

4D clue: Really impresses
Answer: AWES

5D clue: Unfair reputation
Answer: BUMRAP

6D clue: Host of the 2024 Summer Olympics
Answer: PARIS

7D clue: «Warrior-monks who keep peace in the universe,» per George Lucas
Answer: JEDI

How to play more Mini Crosswords

The New York Times Games section offers a large number of online games, but only some of them are free for all to play. You can play the current day’s Mini Crossword for free, but you’ll need a subscription to the Times Games section to play older puzzles from the archives.