I have been travelling the past few weeks, and have not yet had the time to fully digest the recent announcement by Deepmind of how their two new #IMO problem solver engines, AlphaProof and AlphaGeometry2, were able to between them solve 4 of the 6 problems at the most recent International Mathematical Olympiad: deepmind.google/discover/blog/… . But I can record some preliminary impressions.

1. This is great work, shifting once again our expectations of which benchmark challenges are within reach of either #AI-assisted or fully autonomous methods. For instance, IMO level geometry problems are now effectively a solved problem for specialized AI tools; and it seems now that IMO problems that can be readily formalized and with formal proofs that can be located through a reinforcement learning process are now at least somewhat amenable to AI attacks (though currently requiring genuinely significant amounts of compute per problem, and human assistance on the formalization side).

2. There may be side benefits of this approach into making formal mathematics easier to automate, which could in turn facilitate mathematical research methods that contain formal components. In particular, the database of formal proofs generated by this effort could be a useful resource if shared more openly.

3. The approach (based more on reinforcement learning than large language models, somewhat in the spirit of AlphaGo, and heavily emphasizing formal methods) is clever, and makes sense in retrospect. As per the "AI effect" en.wikipedia.org/wiki/AI_effec… , once explained, it does not "feel" like an exhibition of human-like intelligence; but it is still an expansion of the capability of our suite of AI-assisted problem solving tools.

After AI, the next big thing in #tech will be quantum computers that break currently used email #encryption. Fortunately, Tuta Mail is already quantum-safe. 🔒 💪

At Tuta we're not wasting our time on #AI, for a number of ethical reasons 🧵👇

Meta is already using it’s user’s data in the USA, But its brazen attempt to use EU user data and introduce Meta AI won’t pass! 🙅 🧐

Read more here 👉👉tuta.com/blog/noyb-privacy-win…

#MetaAi #Meta #AI #NYOB #Privacyrights #gdpr #GDPR #AI #Tuta #privacy

Generative AI is garbage, exhibit 99:

#ai #math

The factoid that “MS copilot was more successful than 84% of college students at intro level college course tests” isn’t an indication that an LLM is doing something impressive.

It’s an indication that we failed a whole generation of young people by feeding them mediocrity and industrialized info packages instead of meaningful human education.

If you approach your education as a means of turning people into cogs in a machine, of course the machine will be more successful than most of them.

#AI #LLMs #Education

🔥🚨 #AI energy consumption keeps rising so #Google drops claims of being carbon neutral.

At Tuta we are committed to making the web better, more #sustainable and more #private! 💪

👉 tuta.com/sustainability

👉 Read full story here: bloomberg.com/news/articles/20…

Also @Tutanota , as a privacy focused company, why are your comments run on #Reddit, rather than the #Fediverse? Most privacy minded folks don't want their comments being used by AI LLM's etc.

This is way off brand.

@carlschwan among others have already shown how to do it fedistyle, pretty easily, and I'm sure many of the FOSS Fedipeeps here would happily help you out with a quick transition if you asked or gave a few $ to their FOSS project.

#Tuta #Reddit #Privacy #Fediverse #FOSS #AI #LLM

Need a brief and accessible explainer about the actual harms of AI?

Show folks this:

m.youtube.com/watch?v=TtVJ4JDM…

@thejuicemedia hitting it out of the park, yet again.

#ai #corporations #corporatocracy #billionaires #climateCrisis #BigTech #SiliconValley

Honest Government Ad | AI

The Government™ has made an ad about the existential threat that AI poses to humanity, and it’s surprisingly honest and informative👉 Help us keep Government...

^YouTube

The image is a popular internet meme format known as the "Distracted Boyfriend" meme. It shows a man walking with his girlfriend but turning his head to look at another woman passing by. In this version of the meme:

- The man is labeled "MEDIA."
- The woman he is looking at is labeled "REFORM'S 4 MPs."
- The girlfriend, who looks annoyed, is labeled "THE GREENS' 4 MPs."

The meme humorously suggests that the media is more interested in the Reform Party's 4 MPs than in the Greens' 4 MPs. #Blind #AI 👩‍🦯😘 mas.to/@howie319/1127348400461…

Howie319 (@howie319@mas.to)

Attached: 1 image

^mas.to

(YouTube) Lies and hallucinations for the blind? Revolutionary AI smart glasses: #Solos #AirGo #Vision with GPT-4o now unveiled! youtube.com/watch?v=TvSId9u98S… #AI #AR #GPT4o

(AppleVis forum) Solos AirGo Vision: the future of smart glasses for blind and visually impaired users? applevis.com/forum/assistive-t… #blind #a11y

Revolutionary AI Smart Glasses: Solos AirGo Vision with GPT-4o Now Unveiled!

Introducing Groundbreaking Innovation: Solos' AI-powered Smart GlassesSolos, a front-runner in wearable technology, has unveiled a game-changing product that...

^YouTube

GPT 4 hallucination rate is 28.6% on a simple task: citing title, author, and year of publication medium.com/@michaelwood33311/g…

Hallucination rates and reference accuracy of ChatGPT and Bard for systematic reviews: Comparative analysis jmir.org/2024/1/e53164 #AI #LLM

GPT 4 Hallucination Rate is 28.6% on a Simple Task: Citing Title, Author, and Year of Publication

The all-too-common myth of GPT 4 having only a 3% hallucination rate is shattered by a recent study that found GPT 4 has a 28.6% hallucination rate. That’s almost 10x the oft-cited (i.e. over hyped)…

^{Michael Wood (Medium)}

Here's a very calm, well researched video analyzing the facts and data about AI technology progress. And all the stunning tech demos that were faked.

I really like the scientific approach to this and I feel like I'd want to adapt this kind of analytical process to other things in my life.

youtu.be/VctsqOo8wsc?si=XcuufC…

#AiRant #AI #fAIl #tech #science

AI Hype is completely out of control - especially since ChatGPT-4o

Now that ChatGPT4o is out, it's time to revisit my previous thoughts on AI, and discuss what might be happening in the future, and what evidence we have of w...

^YouTube

Upon recommendations, I tried this AI prompt injection game for the first time. I made it to level 7 with no help from the internet!

If you want to donate to me, donate to me on this page.

My website is here where I usually blog. I'm not much of a video person, so I blog and write more than I do video!

One of the selling points of Google’s flagship generative AI models, Gemini 1.5 Pro and 1.5 Flash, is the amount of data they can supposedly process and analyze. But new research suggests that the models aren’t, in fact, very good at those things. “While models like Gemini 1.5 Pro can technically process long contexts, we have seen many cases indicating that the models don’t actually ‘understand’ the content,” Marzena Karpinska, a postdoc at UMass Amherst. TechCrunch has more: flip.it/wW5pRm
#Tech #AI #ArtificialIntelligence #Google #Gemini

Follow @Techcrunch for more breaking tech news and analysis.

And what would these standards be, pray tell, @mozilla ?

Keeping in mind that you are including services that told people to eat rocks or put glue on a pizza?

> Our initial offering will include ChatGPT, Google Gemini, HuggingChat, and Le Chat Mistral, but we will continue adding AI services that meet our standards for quality and user experience.
blog.mozilla.org/en/products/f…

🤡

/via @alxd

#AI

Choose how you want to navigate the web with Firefox

We're launching an opt-in experiment offering access to preferred AI services in Firefox Nightly for improved productivity as you browse.

^{Kristina Bravo (The Mozilla Blog)}

I'm all for the direction generative AI is heading. By using it for good and feeding in quality data, we can harness its potential.

Some believe being pro generative AI means endorsing destruction because they see it as harmful.

Generative AI isn't destructive just because it might replace some jobs. In fact, it creates endless opportunities, including new roles in the job market. Adaptation is key.

#AI #GenerativeAI #Future #ArtificialIntelligence

Wait...What?

microsoft.com/en-us/security/b…

#AI #cybersecurity

AI is an unasked for tech looking for a problem to solve. It uses an incredible amount of energy.

Blockchain was an unasked for tech looking for a problem to solve. It uses an incredible amount of energy.

Pursuit of endless growth is pursuit of a cancer.

Moore's Law is dead and we have reached a happy plateau. We dont need more computing power. We can do so much now. If you need something to chew on, work on making it more efficient. Make the code smaller and faster and lighter.

And less of a power draw.

#AI #Blockchain #SolarPunk

Is anyone interested in a live course learning to build GPT's. It would be talking about what they are and how to build them.

The course would be live and take 2 weeks with 1 class each week and 2 hours for each class. This would be virtual on zoom.

The class would cost around $20.

Who is interested.

#AI

#GPT

#WCAG

#GPT

#OpenAI

#Accessibility

DELL: Putting the IT in SHIT

#AI

This Canadian Member of Parliament asked ChatGPT for a list of capital gains tax rates by country, got a nonsense answer, screenshotted it, and then tweeted the incorrect information.

(He later deleted the tweet.)

#canada #cdnpoli #twitter #x #ai #chatgpt #misinformation

Here's a picture taken with the Meta Smart Glasses, then described with #AI via #BeMyEyes, which has been an invaluable helping-hand, enabling me to know what randomly labeled pictures in my collection are, without having to ask my wife or children.
With image names such as 'photo-296_singular_display_fullPicture.jpg' it's a fun time trying to decipher such randomness sometimes.
This was taken at a podcast recording of Tim Harford's 'Cautionary Tales' and I got to take Jake along with me.

I get the AI hate I really do, but there are just times in life where hate is misplaced. for me, something that is an enabler is such a time.
No, it's not perfect, we're not there yet, but for my usecase it helps more than it hinders.
Your milage will of course, vary.