×
Men Are Buying Hacking Tools to Use Against Their Wives and FriendsThousands of men are members of Telegram groups and channels that advertise and sell hacking and surveillance services that can be used to harass friends, wives and girlfriends, and former partners, new research has uncovered. The findings, from a European nonprofit group, also say that the communities are involved in extensive trading, selling, and promotion of a huge variety of abusive content, including nonconsensual intimate images of women, so-called nudifying services, plus folders of images that sellers claim include child sexual abuse material and depictions of incest and rape.Over six weeks earlier this year, researchers at the algorithmic auditing group AI Forensics analyzed nearly 2.8 million messages sent across 16 Italian and Spanish Telegram communities that are regularly posting abusive content targeting women and girls. More than 24,000 members of the Telegram groups and channels took part in posting 82,723 images, videos, and audio files over the course of the study, the analysis says. Many posts target celebrities and influencers, but men in the groups also frequently victimize women they know.“We tend to forget that most victims are ordinary women who sometimes don’t even know that their pictures are shared or manipulated in these types of channels,” says Silvia Semenzin, a researcher at AI Forensics who previously exposed Italian Telegram channels engaging in similar behavior as far back as 2019. “The majority of this violence is directed towards people who the perpetrators know,” she says, suggesting that Telegram, which has over 1 billion monthly active users, according to company founder Pavel Durov, should be subject to stricter regulation and classed as a “very large online platform” under Europe’s online safety rules.The findings come as Durov is fighting back against Russia’s efforts to block the messaging app in that country, which has long positioned itself as a messaging app that allows free speech but has simultaneously been used by some to share terrorist, sexual abuse, and cybercrime materials. Durov is under criminal investigation in France relating to alleged criminal activity taking place on Telegram, although he has consistently denied the allegations.A Telegram spokesperson tells WIRED that the company removes “millions” of pieces of content per day using “custom AI tools” and has policies in Europe that do not allow the promotion of violence, illegal sexual content including nonconsensual imagery, and other content such as doxing and selling illegal goods and services.Among the extensive types of abusive content and services observed by the AI Forensics researchers were frequent references to the access, publishing, and doxing of women’s private information, sharing their Instagram or TikTok content, as well as references to spying or hacking. “Victims are often named, tagged, and locatable via shared profile links,” the group’s report says.One translated post on Telegram titled “Professional hacking on commission” claimed to be able to give customers “access to phone gallery and extraction of photos and videos,” as well as “anonymous social media hacking.” Another message says: “I hack and recover any type of social media service. I can spy on your partner’s account. Send me a private message.”Across the dataset there were more than 18,000 references to spying or spy content. One post reads: “Hi, do you have the desire to spy on a girl’s gallery? We sell a bot that does it for info DM.” Meanwhile, users were observed asking if people could find phone numbers connected to Instagram accounts and other requests, “who exchanges spy photos and videos?”#Men #Buying #Hacking #Tools #Wives #Friendscrime,privacy,security,cybersecurity,hacking,surveillance,telegram

Men Are Buying Hacking Tools to Use Against Their Wives and Friends

Thousands of men are members of Telegram groups and channels that advertise and sell hacking and surveillance services that can be used to harass friends, wives and girlfriends, and former partners, new research has uncovered. The findings, from a European nonprofit group, also say that the communities are involved in extensive trading, selling, and promotion of a huge variety of abusive content, including nonconsensual intimate images of women, so-called nudifying services, plus folders of images that sellers claim include child sexual abuse material and depictions of incest and rape.

Over six weeks earlier this year, researchers at the algorithmic auditing group AI Forensics analyzed nearly 2.8 million messages sent across 16 Italian and Spanish Telegram communities that are regularly posting abusive content targeting women and girls. More than 24,000 members of the Telegram groups and channels took part in posting 82,723 images, videos, and audio files over the course of the study, the analysis says. Many posts target celebrities and influencers, but men in the groups also frequently victimize women they know.

“We tend to forget that most victims are ordinary women who sometimes don’t even know that their pictures are shared or manipulated in these types of channels,” says Silvia Semenzin, a researcher at AI Forensics who previously exposed Italian Telegram channels engaging in similar behavior as far back as 2019. “The majority of this violence is directed towards people who the perpetrators know,” she says, suggesting that Telegram, which has over 1 billion monthly active users, according to company founder Pavel Durov, should be subject to stricter regulation and classed as a “very large online platform” under Europe’s online safety rules.

The findings come as Durov is fighting back against Russia’s efforts to block the messaging app in that country, which has long positioned itself as a messaging app that allows free speech but has simultaneously been used by some to share terrorist, sexual abuse, and cybercrime materials. Durov is under criminal investigation in France relating to alleged criminal activity taking place on Telegram, although he has consistently denied the allegations.

A Telegram spokesperson tells WIRED that the company removes “millions” of pieces of content per day using “custom AI tools” and has policies in Europe that do not allow the promotion of violence, illegal sexual content including nonconsensual imagery, and other content such as doxing and selling illegal goods and services.

Among the extensive types of abusive content and services observed by the AI Forensics researchers were frequent references to the access, publishing, and doxing of women’s private information, sharing their Instagram or TikTok content, as well as references to spying or hacking. “Victims are often named, tagged, and locatable via shared profile links,” the group’s report says.

One translated post on Telegram titled “Professional hacking on commission” claimed to be able to give customers “access to phone gallery and extraction of photos and videos,” as well as “anonymous social media hacking.” Another message says: “I hack and recover any type of social media service. I can spy on your partner’s account. Send me a private message.”

Across the dataset there were more than 18,000 references to spying or spy content. One post reads: “Hi, do you have the desire to spy on a girl’s gallery? We sell a bot that does it for info DM.” Meanwhile, users were observed asking if people could find phone numbers connected to Instagram accounts and other requests, “who exchanges spy photos and videos?”

#Men #Buying #Hacking #Tools #Wives #Friendscrime,privacy,security,cybersecurity,hacking,surveillance,telegram

Thousands of men are members of Telegram groups and channels that advertise and sell hacking and surveillance services that can be used to harass friends, wives and girlfriends, and former partners, new research has uncovered. The findings, from a European nonprofit group, also say that the communities are involved in extensive trading, selling, and promotion of a huge variety of abusive content, including nonconsensual intimate images of women, so-called nudifying services, plus folders of images that sellers claim include child sexual abuse material and depictions of incest and rape.

Over six weeks earlier this year, researchers at the algorithmic auditing group AI Forensics analyzed nearly 2.8 million messages sent across 16 Italian and Spanish Telegram communities that are regularly posting abusive content targeting women and girls. More than 24,000 members of the Telegram groups and channels took part in posting 82,723 images, videos, and audio files over the course of the study, the analysis says. Many posts target celebrities and influencers, but men in the groups also frequently victimize women they know.

“We tend to forget that most victims are ordinary women who sometimes don’t even know that their pictures are shared or manipulated in these types of channels,” says Silvia Semenzin, a researcher at AI Forensics who previously exposed Italian Telegram channels engaging in similar behavior as far back as 2019. “The majority of this violence is directed towards people who the perpetrators know,” she says, suggesting that Telegram, which has over 1 billion monthly active users, according to company founder Pavel Durov, should be subject to stricter regulation and classed as a “very large online platform” under Europe’s online safety rules.

The findings come as Durov is fighting back against Russia’s efforts to block the messaging app in that country, which has long positioned itself as a messaging app that allows free speech but has simultaneously been used by some to share terrorist, sexual abuse, and cybercrime materials. Durov is under criminal investigation in France relating to alleged criminal activity taking place on Telegram, although he has consistently denied the allegations.

A Telegram spokesperson tells WIRED that the company removes “millions” of pieces of content per day using “custom AI tools” and has policies in Europe that do not allow the promotion of violence, illegal sexual content including nonconsensual imagery, and other content such as doxing and selling illegal goods and services.

Among the extensive types of abusive content and services observed by the AI Forensics researchers were frequent references to the access, publishing, and doxing of women’s private information, sharing their Instagram or TikTok content, as well as references to spying or hacking. “Victims are often named, tagged, and locatable via shared profile links,” the group’s report says.

One translated post on Telegram titled “Professional hacking on commission” claimed to be able to give customers “access to phone gallery and extraction of photos and videos,” as well as “anonymous social media hacking.” Another message says: “I hack and recover any type of social media service. I can spy on your partner’s account. Send me a private message.”

Across the dataset there were more than 18,000 references to spying or spy content. One post reads: “Hi, do you have the desire to spy on a girl’s gallery? We sell a bot that does it for info DM.” Meanwhile, users were observed asking if people could find phone numbers connected to Instagram accounts and other requests, “who exchanges spy photos and videos?”

Source link
#Men #Buying #Hacking #Tools #Wives #Friends

Previous post

Daredevil: Born Again season 2 changes the relationship between Matt Murdock and Bullseye, and that’s a good thing

Next post

Deadspin | NBA roundup: Wolves down Pacers, subsequently clinch playoff berth <div id=""><section id="0" class=" w-full"><div class="xl:container mx-0 !px-4 py-0 pb-4 !mx-0 !px-0"><img src="https://images.deadspin.com/tr:w-900/28679598.jpg" srcset="https://images.deadspin.com/tr:w-900/28679598.jpg" alt="NBA: Minnesota Timberwolves at Indiana Pacers" class="w-full" fetchpriority="high" loading="eager"/><span class="text-0.8 leading-tight">Apr 7, 2026; Indianapolis, Indiana, USA; Minnesota Timberwolves guard Ayo Dosunmu (13) shoots the ball while Indiana Pacers center Jay Huff (32) defends in the first half at Gainbridge Fieldhouse. Mandatory Credit: Trevor Ruszkowski-Imagn Images<!-- --> <!-- --> </span></div></section><section id="section-1"> <p>Ayo Dosunmu scored 24 points as the Minnesota Timberwolves snapped a three-game losing streak with a 124-104 win over the Indiana Pacers on Tuesday in Indianapolis.</p> </section><section id="section-2"> <p>Julius Randle and Bones Hyland added 19 apiece and Naz Reid contributed 17 for Minnesota (47-32), which sits in sixth place in the Western Conference. The Timberwolves finished the night clinching a playoff spot and avoiding the play-in after the Phoenix Suns fell to the Houston Rockets 119-105 later Tuesday.</p> </section><section id="section-3"> <p>Rookie Ethan Thompson posted 17 points for the Pacers (18-61), who took their third straight defeat. Obi Toppin and Jalen Slawson both had 14 points.</p> </section><section id="section-4"> <p>Indiana’s season-long injury crisis continued when forward Kobe Brown, who was in the starting five, was unable to play after halftime with lower back soreness.</p> </section><section id="section-5"> <p>Thunder 123, Lakers 87</p> </section><section id="section-6"> <p>Shai Gilgeous-Alexander scored 25 points in 28 minutes as Oklahoma City demolished the Lakers for the second time in five days, prevailing in Los Angeles.</p> </section><section id="section-7"> <p>The Thunder, by earning their sixth consecutive win and their 18th in 19 games, moved within one win or one San Antonio Spurs loss of clinching the top spot in the Western Conference for the third consecutive season. Chet Holmgren had 15 points and 10 rebounds, while Isaiah Joe and Jared McCain added 18 and 15 points off the bench, respectively.</p> </section><section id="section-8"> <p>The Lakers fell a game behind the Denver Nuggets for the No. 3 spot in the West with their third consecutive loss. They are now tied with Houston after the Rockets beat the Suns. Rui Hachimura led Los Angeles with 15 points on 7-of-10 shooting. No other Los Angeles player scored more than 11.</p> </section><section id="section-9"> <p>Rockets 119, Suns 105</p> </section><section id="section-10"> <p>Kevin Durant had 24 points and five 3-pointers in his return to Phoenix and Houston overcame a 21-point deficit for its largest comeback of the season, producing a victory over the host Suns.</p> </section><section id="section-11"> <p>Amen Thompson logged 22 points, 11 rebounds and eight assists and Jabari Smith Jr. had 20 points and five 3-pointers for the Rockets, who have won seven in a row and nine of 11. The Rockets are tied with the Los Angeles Lakers for fourth place in the West and are one game behind the Denver Nuggets with three to play.</p> </section><section id="section-12"> <p>Devin Booker had 31 points and eight assists and Mark Williams contributed 19 points and eight rebounds for the Suns, who had a 24-0 run in the first quarter to build a 26-5 lead before fading.</p> </section><section id="section-13"> <p>Pelicans 156, Jazz 137</p> </section><section id="section-14"> <p>Jeremiah Fears scored a career-high and franchise-rookie-record 40 points and as New Orleans set a club record for points in a blowout of visiting Utah.</p> </section><section id="section-15"> <p>Jordan Poole added 34 points, Jordan Hawkins put up a season-high 25 and rookie Micah Peavy scored a career-high 20 as the Pelicans ended an eight-game losing streak.</p> </section><section id="section-16"> <p>Kennedy Chandler scored a career-high 31 points and rookie Bez Mbeng added a career-best 26 for the Jazz, who lost their 10th consecutive game.</p> </section><section id="section-17"> <p>Celtics 113, Hornets 102</p> </section><section id="section-18"> <p>Jaylen Brown scored 35 points and grabbed nine rebounds to propel Boston past visiting Charlotte.</p> </section><section id="section-19"> <p>The Celtics received 23 points from Jayson Tatum and 12 apiece from Neemias Queta, Derrick White and Payton Pritchard. Boston has won four straight and 11 of its past 13 games.</p> </section><section id="section-20"> <p>LaMelo Ball tossed in a game-high 36 points for the Hornets, but he failed to score in the fourth quarter. Charlotte was held to 15 points in the final period as its four-game winning streak ended.</p> </section><br/><section id="section-21"> <p>Bulls 129, Wizards 98</p> </section> <section id="section-22"> <p>Rob Dillingham’s career-high 26 points off the bench highlighted a balanced scoring effort as Chicago rolled past host Washington.</p> </section><section id="section-23"> <p>Chicago left little doubt on the way to snapping a seven-game losing streak, pouncing on Washington with a 9-0 run midway the first quarter. Tre Jones put up 20 points, his third effort of 19-plus points in his past five games, and he wound up one assist shy of his first double-double since January. Patrick Williams added 20 points, seven rebounds and six assists.</p> </section><section id="section-24"> <p>Bilal Coulibaly led the Wizards with 19 points. First-year forward Julian Reese posted his fourth double-double in his 10 NBA games, going for 17 points and 11 rebounds.</p> </section><section id="section-25"> <p>Nets 96, Bucks 90</p> </section><section id="section-26"> <p>E.J. Liddell scored 21 points to lead Brooklyn to a victory over visiting Milwaukee.</p> </section><section id="section-27"> <p>It was the second win in a row for Brooklyn and third in its last five contests. Ben Saraf added 19 points while Malachi Smith and Drake Powell each scored 11.</p> </section><section id="section-28"> <p>AJ Green led the Bucks with 20 points on six 3-pointers (6-for-12). Taurean Prince scored 16 points with 11 rebounds while Cormac Ryan added 14. Jericho Sims tallied 12 points, eight boards and six assists, and Ousmane Dieng had 10 points and seven rebounds.</p> </section><section id="section-29"> <p>Clippers 116, Mavericks 103</p> </section><section id="section-30"> <p>Kawhi Leonard scored 34 points and Darius Garland added 22 as Los Angeles improved its play-in tournament seeding possibilities with a victory over Dallas in Inglewood, Calif.</p> </section><section id="section-31"> <p>John Collins scored 12 points and Derrick Jones Jr. added 11 points with 10 rebounds as the eighth-place Clippers moved a full game ahead of the ninth-place Portland Trail Blazers. The Clippers and Blazers will meet Friday at Portland.</p> </section><section id="section-32"> <p>Cooper Flagg scored 25 points and Marvin Bagley III added 21 as the Mavericks fell to 2-9 since March 16. Flagg, 19, coming off consecutive games of 51 and 45 points.</p> </section><section id="section-33"> <p>Raptors 121, Heat 95</p> </section><section id="section-34"> <p>Scottie Barnes scored 25 points and Brandon Ingram added 23, leading Toronto to a victory over visiting Miami.</p> </section><section id="section-35"> <p>Jakob Poeltl had 17 points while RJ Barrett chipped in 16 for the Raptors, who pulled within a game of the fifth-place Atlanta Hawks in the Eastern Conference. Toronto had dropped three of its previous four games.</p> </section><section id="section-36"> <p>The Heat got 24 points from Andrew Wiggins and 14 points apiece from Tyler Herro and Norman Powell. Miami shot just 27.3% (12 of 44) on 3-point attempts while dropping its ninth game in 12 tries.</p> </section><section id="section-37"> <p>Warriors 110, Kings 105</p> </section><section id="section-38"> <p>Stephen Curry capped his 17-point performance with a game-tying 3-pointer, then turned his own miss into an offensive rebound and assist on Brandin Podziemski’s go-ahead trey as Golden State outlasted Sacramento in San Francisco.</p> </section><section id="section-39"> <p>De’Anthony Melton had a game-high 21 points and Podziemski finished with 20 for the Warriors, who ended a four-game losing streak while resting Kristaps Porzingis on the first night of a four-games-in-six-days sequence.</p> </section><section id="section-40"> <p>Killian Hayes led a balanced attack with 18 points off the bench for the Kings, who dropped their second in a row.</p> </section><section id="section-41"> <p>–Field Level Media</p> </section></div> #Deadspin #NBA #roundup #Wolves #Pacers #subsequently #clinch #playoff #berth

Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a $20 million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.

While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either.

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

Techcrunch event

San Francisco, CA | October 13-15, 2026

So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

I can’t help rooting for tiny open source AI model maker Arcee | TechCrunch
Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a  million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.







While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either. 

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

	
		
		Techcrunch event
		
			
			
									San Francisco, CA
													|
													October 13-15, 2026
							
			
		
	


So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.








#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai
Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.

#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai">I can’t help rooting for tiny open source AI model maker Arcee | TechCrunch
Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a  million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.







While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either. 

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

	
		
		Techcrunch event
		
			
			
									San Francisco, CA
													|
													October 13-15, 2026
							
			
		
	


So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.








#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai

Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a $20 million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.

While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either.

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

Techcrunch event

San Francisco, CA | October 13-15, 2026

So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

I can’t help rooting for tiny open source AI model maker Arcee | TechCrunch
Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a  million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.







While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either. 

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

	
		
		Techcrunch event
		
			
			
									San Francisco, CA
													|
													October 13-15, 2026
							
			
		
	


So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.








#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai
Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.

#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai">I can’t help rooting for tiny open source AI model maker Arcee | TechCrunch

Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a $20 million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.

While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either.

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

Techcrunch event

San Francisco, CA | October 13-15, 2026

So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

I can’t help rooting for tiny open source AI model maker Arcee | TechCrunch
Arcee, a tiny 26-person U.S. startup that built a massive, 400B-parameter open source LLM on a  million shoestring budget, has released its new reasoning model. Arcee calls the model Trinity Large Thinking — and it’s the most capable open-weight model “ever released by a non-Chinese company,” claims CEO Mark McQuade to TechCrunch.

As that comment implies, Arcee has a goal that I can’t help but root for: It wants to give U.S. and Western companies a model that gives them no reason to use a Chinese-based one.







While Chinese models are extremely capable, they are perceived as risky, putting power, and perhaps data, into the hands of a government that doesn’t share all of the Western world’s ideals.

With Arcee, companies can download the model, train it to their own needs, and use it on premises. Companies can also use Arcee’s cloud-hosted version, accessible via API.

While Arcee’s models are not outperforming the closed source models from the big labs like Anthropic or OpenAI, they’re not being held hostage by the whims of those giants, either. 

For instance, Claude, with its exceptional abilities to code, has been a popular choice for users of open source AI agent tool OpenClaw. But Anthropic pulled the rug out from them last week when it told users that their Anthropic subscriptions will no longer cover OpenClaw usage — they will have to pay additionally for that. (In February, OpenClaw creator Peter Steinberger said he was joining Anthropic’s biggest rival, OpenAI.)

In contrast, McQuade proudly points to data from OpenRouter that says it has become one of the top models used with OpenClaw.

	
		
		Techcrunch event
		
			
			
									San Francisco, CA
													|
													October 13-15, 2026
							
			
		
	


So, how good is Trinity Large Thinking? It is comparable to some of the other top open source models, according to the benchmark results it shared with TechCrunch.

Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.








#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai
Arcee Trinity large thinking BenchmarksImage Credits:Arcee / Arcee

As we previously reported, it is not a head-to-head threat to the big cheese among U.S.-built open models: Meta’s Llama 4. But it also doesn’t have the odd, not-really open source license issues of Meta’s model. All of Arcee’s Trinity models are released under the gold standard for OS licenses, Apache 2.0.

Just to be clear, there are also countless other U.S. startups offering open source models and, as a fan of the ingenuity of startups, I’m rooting for them, too.

#rooting #tiny #open #source #model #maker #Arcee #TechCrunchArcee AI,chinese ai,open source ai

Post Comment