Andrew Tarantola
@terrortola.bsky.social
I deliver mail
Check out some of my best writing from a past life at AndrewTarantola.com.
Check out some of my best writing from a past life at AndrewTarantola.com.
After declining management’s offer to work for the site full time -- but as a freelancer, with no medical benefits or 401k, while earning less than I made as an entry-level editorial assistant for Gawker in 2010 -- today is my last day with Digital Trends. Thank Christ.
February 28, 2025 at 4:25 PM
After declining management’s offer to work for the site full time -- but as a freelancer, with no medical benefits or 401k, while earning less than I made as an entry-level editorial assistant for Gawker in 2010 -- today is my last day with Digital Trends. Thank Christ.
Someone didn't get a bite of hotdog and is now Very Upset
February 18, 2025 at 10:40 PM
Someone didn't get a bite of hotdog and is now Very Upset
Just realized that each one of my pets have their own tragic origin stories.
February 17, 2025 at 7:18 PM
Just realized that each one of my pets have their own tragic origin stories.
Being a freelancer is wild. Every day I wake up to a new panic attack as to whether the current company's clownshow COO decides that today is the day that I lose my job or not.
February 14, 2025 at 5:56 PM
Being a freelancer is wild. Every day I wake up to a new panic attack as to whether the current company's clownshow COO decides that today is the day that I lose my job or not.
Have my fingers crossed that I get a call back on the part time bar back position at my local casino that I applied for, in case you're wondering how my time in tech media is going.
February 13, 2025 at 6:58 PM
Have my fingers crossed that I get a call back on the part time bar back position at my local casino that I applied for, in case you're wondering how my time in tech media is going.
Interesting, of the half dozen houses in my neighborhood that flew Trump/Pence "Finish the Wall" flags for most of 2024, all but one have removed them since the inauguration. Not seeing anyone wearing Trump branded clothes or red hats in town either (pop 2500, county votes overwhelmingly GOP).
February 8, 2025 at 11:57 PM
Interesting, of the half dozen houses in my neighborhood that flew Trump/Pence "Finish the Wall" flags for most of 2024, all but one have removed them since the inauguration. Not seeing anyone wearing Trump branded clothes or red hats in town either (pop 2500, county votes overwhelmingly GOP).
Reposted by Andrew Tarantola
Lawmakers in South Korea climbing over fences
Lawmakers in America stopping because a security guard said “nah”
Lawmakers in America stopping because a security guard said “nah”
February 7, 2025 at 1:42 AM
Lawmakers in South Korea climbing over fences
Lawmakers in America stopping because a security guard said “nah”
Lawmakers in America stopping because a security guard said “nah”
weird how this keeps happening
AI researchers at Stanford and the University of Washington were able to train an AI “reasoning” model for under $50 in cloud compute credits, according to a new research paper released last Friday.
Read more here: tcrn.ch/4aMr0vS
Read more here: tcrn.ch/4aMr0vS
February 5, 2025 at 11:46 PM
weird how this keeps happening
That rural folks aren't as tuned in & informed as urbanites are of this news cycle is a myth.
Just went to my local Walmart to stock up before the tariffs hit and apparently so too was everybody else in the county.
Just went to my local Walmart to stock up before the tariffs hit and apparently so too was everybody else in the county.
January 31, 2025 at 11:23 PM
That rural folks aren't as tuned in & informed as urbanites are of this news cycle is a myth.
Just went to my local Walmart to stock up before the tariffs hit and apparently so too was everybody else in the county.
Just went to my local Walmart to stock up before the tariffs hit and apparently so too was everybody else in the county.
Turns out, you can disable Google's AI search results by dropping an f-bomb alongside your search query.
Coincidentally, the same trick works for getting a live rep on the line when you're stuck navigating an automated phone system. Just start swearing until a person picks up.
Coincidentally, the same trick works for getting a live rep on the line when you're stuck navigating an automated phone system. Just start swearing until a person picks up.
“Just give me the f***ing links!”—Cursing disables Google’s AI overviews
The latest trick to stop those annoying AI answers is also the most cathartic.
arstechnica.com
January 31, 2025 at 5:52 PM
Turns out, you can disable Google's AI search results by dropping an f-bomb alongside your search query.
Coincidentally, the same trick works for getting a live rep on the line when you're stuck navigating an automated phone system. Just start swearing until a person picks up.
Coincidentally, the same trick works for getting a live rep on the line when you're stuck navigating an automated phone system. Just start swearing until a person picks up.
I used to manage auto repair shops — did so for nearly a decade — and can assure you that your mechanic is going to absolutely hate this.
Google’s “Ask for Me” will have an AI schedule your next oil change
Google released a new experimental AI agent that can set up nail salon and automotive shop appointments nearly automatically.
www.digitaltrends.com
January 30, 2025 at 9:48 PM
I used to manage auto repair shops — did so for nearly a decade — and can assure you that your mechanic is going to absolutely hate this.
Berkeley researchers reproduced the DeepSeek R1-Zero’s core technologies and taught it to play Countdown (like the game show) for just $30.
"Through [reinforcement learning], the 3B base LM develops self-verification and search abilities all on its own"
www.tomshardware.com/tech-industr...
"Through [reinforcement learning], the 3B base LM develops self-verification and search abilities all on its own"
www.tomshardware.com/tech-industr...
AI research team claims to reproduce DeepSeek core technologies for $30 — relatively small R1-Zero model has remarkable problem-solving abilities
It's cheap and powerful.
www.tomshardware.com
January 29, 2025 at 7:11 PM
Berkeley researchers reproduced the DeepSeek R1-Zero’s core technologies and taught it to play Countdown (like the game show) for just $30.
"Through [reinforcement learning], the 3B base LM develops self-verification and search abilities all on its own"
www.tomshardware.com/tech-industr...
"Through [reinforcement learning], the 3B base LM develops self-verification and search abilities all on its own"
www.tomshardware.com/tech-industr...
Reposted by Andrew Tarantola
OpenAI is furious that DeepSeek might have stolen all of the data OpenAI stole from all of us
www.404media.co/openai-furio...
www.404media.co/openai-furio...
OpenAI Furious DeepSeek Might Have Stolen All the Data OpenAI Stole From Us
OpenAI shocked that an AI company would train on someone else's data without permission or compensation.
www.404media.co
January 29, 2025 at 2:56 PM
OpenAI is furious that DeepSeek might have stolen all of the data OpenAI stole from all of us
www.404media.co/openai-furio...
www.404media.co/openai-furio...
Meanwhile, in Germany
‘Groundbreaking’: scientists develop patch that can repair damaged hearts
‘Groundbreaking’: scientists develop patch that can repair damaged hearts
Cells taken from blood and ‘reprogrammed’ into heart muscle cells may help patients with heart failure
Damaged hearts can literally be patched up to help them work, say researchers, in what has been hailed as a groundbreaking development for people with…
www.theguardian.com
January 29, 2025 at 5:23 PM
Meanwhile, in Germany
Reposted by Andrew Tarantola
People often claim they know when ChatGPT wrote something, but are they as accurate as they think?
Turns out that while general population is unreliable, those who frequently use ChatGPT for writing tasks can spot even "humanized" AI-generated text with near-perfect accuracy 🎯
Turns out that while general population is unreliable, those who frequently use ChatGPT for writing tasks can spot even "humanized" AI-generated text with near-perfect accuracy 🎯
January 28, 2025 at 2:55 PM
People often claim they know when ChatGPT wrote something, but are they as accurate as they think?
Turns out that while general population is unreliable, those who frequently use ChatGPT for writing tasks can spot even "humanized" AI-generated text with near-perfect accuracy 🎯
Turns out that while general population is unreliable, those who frequently use ChatGPT for writing tasks can spot even "humanized" AI-generated text with near-perfect accuracy 🎯
Reposted by Andrew Tarantola
no reason australyear.blogspot.com/2008/05/re-o...
January 28, 2025 at 9:40 PM
no reason australyear.blogspot.com/2008/05/re-o...
time for some self-care nachos
January 28, 2025 at 8:07 PM
time for some self-care nachos
Reposted by Andrew Tarantola
please vet information and links before sharing.
it's very easy to see something that looks like it's going to help people and share it. The reality is that people will use this time to share scams and false information with dire consequences.
it's very easy to see something that looks like it's going to help people and share it. The reality is that people will use this time to share scams and false information with dire consequences.
January 28, 2025 at 5:48 PM
please vet information and links before sharing.
it's very easy to see something that looks like it's going to help people and share it. The reality is that people will use this time to share scams and false information with dire consequences.
it's very easy to see something that looks like it's going to help people and share it. The reality is that people will use this time to share scams and false information with dire consequences.
Both Sen Padilla's and Sen Schiff's DC offices went straight to voicemail because their lines were all busy with other calls, I assume that's a good thing?
![YOU: Hi, my name is [NAME]. I’m calling from [ZIP CODE].
YOU: I am calling on [HOUSE MEMBER/SENATOR] to speak out against President Trump’s actions on Monday in freezing all federal aid. This is blatantly illegal; the Executive Branch can’t unilaterally decide to ignore laws that were passed by Congress, especially when it comes to financial aid.
YOU: MOC must take action, and clarify how they will stop the freeze and ensure that the millions who are directly impacted by financial aid will not be harmed.
YOU: [OPTIONAL COMMENT]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:vdabazxnghujbebpi2ektqgi/bafkreia6t55htxgj6v4nxv5uys3mklsp7lgba2qajqxb6c542lp2gf4gq4)
January 28, 2025 at 6:05 PM
Both Sen Padilla's and Sen Schiff's DC offices went straight to voicemail because their lines were all busy with other calls, I assume that's a good thing?
Reposted by Andrew Tarantola
i think a lot of people on here should be logging off and learning how to win a fistfight and safely apply a tourniquet bc your posts suck ass and i can't imagine you're any better in real life
January 28, 2025 at 12:47 PM
i think a lot of people on here should be logging off and learning how to win a fistfight and safely apply a tourniquet bc your posts suck ass and i can't imagine you're any better in real life
Wait, he can just fucking do that?
January 28, 2025 at 5:03 PM
Wait, he can just fucking do that?
Just ate my first tuna melt. Hot sandwich fish is not a favorite.
January 25, 2025 at 1:40 AM
Just ate my first tuna melt. Hot sandwich fish is not a favorite.
You ever wonder how much pet hair you inadvertently eat in a given week?
January 24, 2025 at 5:30 PM
You ever wonder how much pet hair you inadvertently eat in a given week?
Reposted by Andrew Tarantola
Chinese start-up DeepSeek has released R1 - a 'reasoning model' that's as good as OpenAI's o1 but 1/30th of the price. Incorporating other AI techniques into LLMs is boosting their ability to tackle complex tasks, which could someday prove a huge boon to research 🤖🧪 www.nature.com/articles/d41...
China’s cheap, open AI model DeepSeek thrills scientists
DeepSeek-R1 performs reasoning tasks at the same level as OpenAI’s o1 — and is open for researchers to examine.
www.nature.com
January 23, 2025 at 4:54 PM
Chinese start-up DeepSeek has released R1 - a 'reasoning model' that's as good as OpenAI's o1 but 1/30th of the price. Incorporating other AI techniques into LLMs is boosting their ability to tackle complex tasks, which could someday prove a huge boon to research 🤖🧪 www.nature.com/articles/d41...