Technologies

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

ChatGPT Plus subscribers can try it out now.

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

The new 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 to 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 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.

OpenAI has made its GPT technology available to developers for years, but ChatGPT, which debuted in November, offered an easy interface that yielded an explosion of interest, experimentation and worry about the downsides of the technology. ChatGPT is free, but it 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.

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.

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, 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.

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.

Technologies

Today’s NYT Connections Hints, Answers and Help for Dec. 24, #927

Here are some hints and the answers for the NYT Connections puzzle for Dec. 24 #927

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

Today’s NYT Connections puzzle is kind of tough. Ooh, that purple category! Once again, you’ll need to look inside words for hidden words. Read on for clues and today’s Connections answers.

The Times has a Connections Bot, like the one for Wordle. Go there after you play to receive a numeric score and to have the program analyze your answers. Players who are registered with the Times Games section can now nerd out by following their progress, including the number of puzzles completed, win rate, number of times they nabbed a perfect score and their win streak.

Read more: Hints, Tips and Strategies to Help You Win at NYT Connections Every Time

Hints for today’s Connections groups

Here are four hints for the groupings in today’s Connections puzzle, ranked from the easiest yellow group to the tough (and sometimes bizarre) purple group.

Yellow group hint: Cash out.

Green group hint: Chomp

Blue group hint: Walleye and salmon.

Purple group hint: Make a musical sound, with a twist.

Answers for today’s Connections groups

Yellow group: Slang for money.

Green group: Masticate.

Blue group: Fish.

Purple group: Ways to vocalize musically plus a letter.

Read more: Wordle Cheat Sheet: Here Are the Most Popular Letters Used in English Words

What are today’s Connections answers?

The yellow words in today’s Connections

The theme is slang for money. The four answers are bacon, bread, cheese and paper.

The green words in today’s Connections

The theme is masticate. The four answers are bite, champ, chew and munch.

The blue words in today’s Connections

The theme is fish. The four answers are char, pollock, sole and tang.

The purple words in today’s Connections

The theme is ways to vocalize musically plus a letter. The four answers are hump (hum), rapt (rap), singe (sing) and whistler (whistle).

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

Toughest Connections puzzles

We’ve made a note of some of the toughest Connections puzzles so far. Maybe they’ll help you see patterns in future puzzles.

#5: Included «things you can set,» such as mood, record, table and volleyball.

#4: Included «one in a dozen,» such as egg, juror, month and rose.

#3: Included «streets on screen,» such as Elm, Fear, Jump and Sesame.

#2: Included «power ___» such as nap, plant, Ranger and trip.

#1: Included «things that can run,» such as candidate, faucet, mascara and nose.

Technologies

Today’s NYT Mini Crossword Answers for Wednesday, Dec. 24

Here are the answers for The New York Times Mini Crossword for Dec. 24.

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.

Need some help with today’s Mini Crossword? I’m Irish-American, but yet 6-Down, which involves Ireland, stumped me at first. Read on for all the answers.. And if you could use some hints and guidance for daily solving, check out our Mini Crossword tips.

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 to those Mini Crossword clues and answers.

Mini across clues and answers

1A clue: Wordle or Boggle
Answer: GAME

5A clue: Big Newton
Answer: ISAAC

7A clue: Specialized vocabulary
Answer: LINGO

8A clue: «See you in a bit!»
Answer: LATER

9A clue: Tone of many internet comments
Answer: SNARK

Mini down clues and answers

1D clue: Sharks use them to breathe
Answer: GILLS

2D clue: From Singapore or South Korea, say
Answer: ASIAN

3D clue: Large ocean ray
Answer: MANTA

4D clue: ___ beaver
Answer: EAGER

6D clue: Second-largest city in the Republic of Ireland, after Dublin
Answer: CORK

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

Technologies

Quadrantids Is a Short but Sweet Meteor Shower Just After New Year’s. How to See It

This meteor shower has one of the most active peaks, but it doesn’t last for very long.

The Quadrantids has the potential to be one of the most active meteor showers of the year, and skygazers won’t have long to wait to see it. The annual shower is predicted to reach maximum intensity on Jan. 3. And with a display that can rival Perseids, Quadrantids could be worth braving the cold to see it.

Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.

The show officially begins on Dec. 28 and lasts until Jan. 12, according to the American Meteor Society. Quadrantids is scheduled to peak on Jan. 2-3, when it may produce upwards of 125 meteors per hour. This matches Perseids and other larger meteor showers on a per-hour rate, but Quadrantids also has one of the shortest peaks at just 6 hours, so it rarely produces as many meteors overall as the other big ones.

The meteor shower comes to Earth courtesy of the 2003 EH1 asteroid, which is notable because most meteor showers are fed from comets, not asteroids. Per NASA, 2003 EH1 is a near-Earth asteroid that orbits the sun once every five and a half years. Science posits that 2003 EH1 was a comet in a past life, but too many trips around the sun stripped it of its ice, leaving only its rocky core. The Earth runs through EH1’s orbital debris every January, which results in the Quadrantids meteor shower.

How and where to see Quadrantids

Quadrantids is named for the constellation where its meteors appear to originate, a point known as the radiant. This presents another oddity, as the shower originates from the constellation Quadrans Muralis. This constellation ceased to be recognized as an official constellation in the 1920s and isn’t available on most publicly accessible sky maps.

For the modern skygazer, you’ll instead need to find the Bootes and Draco constellations, both of which contain stars that were once a part of the Quadrans Muralis. Draco will be easier to find after sunset on the evening of Jan. 2, and will be just above the horizon in the northern sky. Bootes orbits around Draco, but will remain under the horizon until just after 1 a.m. local time in the northeastern sky. From that point forward, both will sit in the northeastern part of the sky until sunrise. You’ll want to point your chair in that direction and stay there to see meteors.

As the American Meteor Society notes, Quadrantids has a short but active peak, lasting around 6 hours. The peak is expected to start around 4 p.m. ET and last well into the evening. NASA predicts the meteor shower to start one day later on Jan. 3-4, so if you don’t see any on the evening of Jan. 2, try again on Jan. 3.

To get the best results, the standard space viewing tips apply. You’ll want to get as far away from the city and suburbs as possible to reduce light pollution. Since it’ll be so cold outside, dress warmly and abstain from alcoholic beverages, as they can affect your body temperature. You won’t need any binoculars or telescopes, and the reduced field of view may actually impact your ability to see meteors.

The bad news is that either way, the Quadrantids meteor shower coincides almost perfectly with January’s Wolf Moon, which also happens to be a supermoon. This will introduce quite a lot of light pollution, which will likely drown out all but the brightest meteors. So, while it may have a peak of over 100 meteors per hour, both NASA and the AMS agree that the more realistic expectation is 10 or so bright meteors per hour.