Phrasing requests in a certain way — meanly or nicely — can yield better results with chatbots like ChatGPT than prompting in a more neutral tone.
So what’s the deal with emotive prompts?
Nouha Dziri, a research scientist at the Allen Institute for AI, theorizes that emotive prompts essentially “manipulate” a model’s underlying probability mechanisms.
Why is it so trivial to defeat safeguards with emotive prompts?
Another reason could be a mismatch between a model’s general training data and its “safety” training datasets, Dziri says — i.e.
Well, it seems the EV market is a tad more cutthroat than anticipated, with pricing pressure mounting like the suspense in a bad thriller movie.
Most interesting startup stories this weekStaying in the world of automotive, there’s been a lot of movement from the avant-garde of EV manufacturers.
These price adjustments from Ford and Lucid come hot on the heels of EV industry poster child Tesla’s price reductions, suggesting that the EV market is maturing and that customers are becoming more price conscious.
Despite all their rage, they’re still just a rat in a Faraday cage: Struggling EV startup Faraday Future owes the landlord of its Los Angeles headquarters nearly $1 million after missing the last two months’ rent.
This week’s big trend: The AI train keeps rumblin’ alongOpenAI has introduced a new generative AI model named Sora, capable of creating videos from text descriptions or still images.
Stability has announced Stable Diffusion 3, the latest and most powerful version of the company’s image-generating AI model.
Sora, OpenAI’s impressive video generator, apparently works on similar principles (Will Peebles, co-author of the paper, went on to co-lead the Sora project).
(Anthropic, for its part, has not focused on image or video generation publicly, so it isn’t really part of this conversation.)
Stable Diffusion seems to want to be the white label generative AI that you can’t do without, rather than the boutique generative AI you aren’t sure you need.
Interestingly, the company has put safety front and center in its announcement, stating:We have taken and continue to take reasonable steps to prevent the misuse of Stable Diffusion 3 by bad actors.
Facebook, Instagram, Snap, YouTube, and other social networking companies offer programs to connect creators with brands, and now Spotify is doing the same.
The company announced the launch of AUX, its new in-house “music advisory agency” for brands.
While not necessarily a creator marketplace, the program has a similar aim — it will facilitate connections between brands and emerging artists for various campaigns benefitting both parties.
For Spotify, AUX represents another source of income, as well, as the company says brands can pay Spotify to leverage the new service.
“Spotify is always looking for ways to leverage our music ecosystem to deepen the connections between artists, brands, and fans,” said Jeremy Erlich, VP, Head of Music Content at Spotify, in a statement.
Investing app Grifin today officially launched its anticipated investing model called “Adaptive Investing,” which enables you to automatically invest in your favorite brands that you frequently shop from.
“Investing, and even having a healthy positive relationship with money, is an incredibly difficult thing to do and achieve,” co-founder Aaron Froug tells TechCrunch.
It also introduces a “Secret Cash” function, allowing for non-public purchases and putting more money away as cash for their future.
“I’ve been personally using our app for a little over two years and I’ve invested in 115 unique companies,” he notes.
Additionally, Grifin is planning a redesign of its app, which will include a premium version as well as an AI chatbot to help people learn how to invest.
Massive training data sets are the gateway to powerful AI models — but often, also those models’ downfall.
Morcos’ company, DatologyAI, builds tooling to automatically curate data sets like those used to train OpenAI’s ChatGPT, Google’s Gemini and other like GenAI models.
“However, not all data are created equal, and some training data are vastly more useful than others.
History has shown automated data curation doesn’t always work as intended, however sophisticated the method — or diverse the data.
The largest vendors today, from AWS to Google to OpenAI, rely on teams of human experts and (sometimes underpaid) annotators to shape and refine their training data sets.
So in response, Google — thousands of jobs lighter than it was last fiscal quarter — is funneling investments toward AI safety.
This morning, Google DeepMind, the AI R&D division behind Gemini and many of Google’s more recent GenAI projects, announced the formation of a new organization, AI Safety and Alignment — made up of existing teams working on AI safety but also broadened to encompass new, specialized cohorts of GenAI researchers and engineers.
But it did reveal that AI Safety and Alignment will include a new team focused on safety around artificial general intelligence (AGI), or hypothetical systems that can perform any task a human can.
The AI Safety and Alignment organization’s other teams are responsible for developing and incorporating concrete safeguards into Google’s Gemini models, current and in-development.
One might assume issues as grave as AGI safety — and the longer-term risks the AI Safety and Alignment organization intends to study, including preventing AI in “aiding terrorism” and “destabilizing society” — require a director’s full-time attention.
Ford is cutting prices of its all-electric 2023 Mustang Mach-E by has much as $8,100 as the automaker attempts to rid itself of inventory and compete with Tesla and its increasingly cheaper EVs.
Total market share of new EV sales has grown, reaching nearly 8% in U.S. in 2023.
Ford confirmed with TechCrunch the price cuts, which are only for model year 2023 Mustang Mach-E vehicles and range between $3,100 and $8,100.
“The Mustang Mach-E is America’s No.2 EV SUV in 2023 and Ford is America’s No.2 EV brand,” Ford spokesperson Marty Günsberg wrote in an emailed statement.
Tesla shipped a record number of electric vehicles in the fourth quarter, which helped it reach 1.81 million deliveries in 2023.
There’s a new generative AI startup based in Paris.
Bioptimus will leverage this unique data set to train its foundational model.
Creating new AI models is such a daunting task that creating a separate entity made more sense.
“As a ‘pure player’ in foundational models, Bioptimus is better set up to do this.”The startup has also signed a partnership with Amazon Web Services.
Now that Bioptimus is well funded, it’s time to work on the AI model and see what the biotech research community can do with it.
OpenAI releases Sora, a credit score–based dating app launches and an anti-Tesla ad comes under fireWelcome, folks, to Week in Review (WiR), TechCrunch’s regular newsletter covering noteworthy happenings in the tech industry.
This week, OpenAI stunned the blogosphere with the release of Sora, a new AI model that can generate videos in impressively high fidelity.
Elsewhere, startup Score released a dating app exclusive to people with good to excellent credit scores.
Google upgrades Gemini: Google expanded the range of its Gemini AI models available to developers across its platforms.
And it’s previewed a new Gemini model capable of analyzing whole books, hours-long audio and hour-long videos.
