Anthropic, the AI startup backed by Google and with substantial venture capital, has just introduced the latest iteration of its GenAI technology: Claude 3. This announcement marks a significant advancement in AI capabilities, positioning Claude 3 as a formidable competitor even against OpenAI’s GPT-4.

Advanced Capabilities

According TechCrunch, “Claude 3, as Anthropic’s new GenAI is called, is a family of models — Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus, Opus being the most powerful. All show “increased capabilities” in analysis and forecasting, Anthropic claims, as well as enhanced performance on specific benchmarks versus models like ChatGPT and GPT-4 (but not GPT-4 Turbo) and Google’s Gemini 1.0 Ultra (but not Gemini 1.5 Pro).”

Multimodal Functionality

One notable feature of Claude 3 is its multimodal functionality, enabling it to analyze both text and images. This capability, like some iterations of GPT-4 and Gemini, allows Claude 3 to process various visual data such as, “…photos, charts, graphs and technical diagrams, drawing from PDFs, slideshows and other document types.” TechCrunch went further to note, “In a step one better than some GenAI rivals, Claude 3 can analyze multiple images in a single request (up to a maximum of 20). This allows it to compare and contrast images, notes Anthropic.” However, Anthropic has imposed limits on image processing to address ethical concerns, “Anthropic has disabled the models from identifying people…”

Claude 3’s Limitations

While Claude 3 showcases remarkable advancements, it’s not without limitations. TechCrunch reported that, “…the company admits that Claude 3 is prone to making mistakes with “low-quality” images (under 200 pixels) and struggles with tasks involving spatial reasoning (e.g. reading an analog clock face) and object counting (Claude 3 can’t give exact counts of objects in images).” Anthropic promises frequent updates to Claude 3, aiming to enhance its capabilities and address existing limitations. These updates will include improvements in following multi-step instructions, structured output generation, and multilingual support, making Claude 3 more responsive and adaptable to user needs.

As Anthropic continues to innovate and expand their offerings, the company remains dedicated to fostering a transparent and responsible approach to AI development. With substantial backing and a clear roadmap for future enhancements, Anthropic is poised to share the future of AI-driven solutions and pave the way for transformative advancements in various domains.

If you’re looking to take your career to the next level be sure to check out our IT page.

author avatar
Green Key
Tags: , , ,

Studies, Surveys Tell the Story of COVID’s Cybersecurity Risks

When the COVID-19 lockdown hit, companies worldwide transitioned millions of employees from working in offices to working at home. There were bumps to be sure, but from an IT perspective the process generally went smoothly.

What’s happened since then is enough to keep IT security professionals up at night.

“Once the transition was complete,” says an article on CSO.com, “Organizations found their attack surface had changed immensely and threat actors attempted to seize upon the opportunity. Phishing, brute-force and malware attacks surged while the number of endpoints connecting to corporate networks ballooned.”

We blogged about this subject a few weeks ago when a survey of IT leaders reported that 41% of them had experienced more security attacks than ever.

In light of the collection of surveys and studies in the CSO.com article, that now looks like an understatement.

Though the study we referenced in our post said in the early days of the lockdown companies were spending an extra $15 billion a week on IT, CSO cites a study that helps explain why: 66% of organizations had no pandemic preparedness plan in place. Others, including those that did, failed to account for the sheer scale of having every employee working remotely.

Infoblox’s COVID-19 Challenges for the Borderless Enterprise report said 38% of organizations shifted funds from cybersecurity to provide for remote worker access. 46%, however, shifted IT resources to shore up the security of their networks. Another study cited by CSO.com tells us that 60% of organizations that adopted work-from-home technology accelerated or bypassed their normal privacy/security reviews.

Consequently says CSO.com, chief information security officers “should go back and ensure that any checks that were skipped or accelerated have been redone to ensure all the risks have been accounted for.”

The article cites Zoom’s security issues as one example of a remote tool that was quickly adopted by many without considering security.

The most worrisome part of the article by CSO editor Dan Swinhoe cites a baker’s dozen of studies, surveys and reports of cyberattacks skyrocketing during the lockdown with many continuing unabated since. Here’s a sample:

  • Supply chain attacks rose 38% since the start of the pandemic;
  • Phishing incidents rose 220% at the height of the pandemic;
  • Ransomware attacks spiked more than 100%;
  • Insider-threats increased 27%;
  • RDP brute-force attacks (attempts to remotely control a computer or computer system) grew 400%.

With the majority of companies expecting more employees than ever to work from home even when the pandemic ends, a PwC Insights Survey found 96% of organizations saying they are adjusting their cybersecurity strategy due to COVID-19. 50% said cybersecurity and privacy will be baked into every business decision or plan.

“This focus on security,” observes CSO, “Should provide CISOs with more influence at the most senior levels of the business.”

Photo by Jefferson Santos on Unsplash


author avatar
Green Key

Leave a Reply

Your email address will not be published. Required fields are marked *

This field is required.

This field is required.