Cybersecurity workers are burned out and its having an impact, Google may be losing its supremacy in answering tech support issues and executives say they’d rather have a robot making key decisions.
These stories and more on Hashtag Trending, for Monday, April 24th. I’m your host Jim Love, CIO of ITWC – IT World Canada and TechNewsDay in the US.
The recent wave of tech layoffs is affecting cybersecurity workers, and not in a good way. It’s leaving them exhausted, overextended, and many are eager to quit their jobs.
A survey done by pen testing firm Cobalt asserts that 77 per cent of security professionals say their departments have faced layoffs in the last six months. Another study done by HackerOne says that 40 per cent of companies plan to make even more security headcount in the next 12 months.
This has left workers strained, with 61 per cent of cyber workers surveyed reporting burnout according to the Cobalt survey.
Shrinking security budgets and decreasing worker headcounts are leaving companies with more unpatched, hackable flaws in their systems.
Half of workers affected by layoffs now want to quit their jobs, and experts say organizations will need to get creative to stave off burnout and keep their networks secure as they battle uncertain economic conditions.
Catherine Wong, Cobalt’s Chief Strategy Officer told Axios, Even if someone doesn’t quit, but they want to quit, they’re not going to be doing their job at 100 per cent if they’re feeling so overwhelmed and so burned out.” She added, “A person in that mental state is not going to be in a position to do an extraordinary job.”
Auto-GPT is the latest in hot AI offerings and it’s been exploding on social media. As we noted in our story last week, this is an open-source app created by a game developer and it’s capable of using either GPT-3.5 or even GPT-4. But the big difference is that where ChatGPT requires prompting AutoGPT allows you to set a goal, and it will develop its own prompts driving to that goal.
Auto-GPT also has an ability to interact with other apps, software and services both online and local, like web browsers and word processors. So, where GPT-3.5 stalls because it doesn’t have data past 2021, AutoGPT tells you that it needs to access the web – and it will do it.
The user sets the role – what or who is the person that Auto-GPT is emulating. So, you could ask it to take the role of an office manager who is charged with finding the best printer.
Then you give it an objective – find the best printer from all of those available on the market – and give me a list of the top 5 choices along with the pluses and minuses of each.
AutoGPT will go and to that task, developing its own prompts along the way. It will even look at areas where it is weak and how it can improve.
You don’t even need to read the response – it has speech synthesis that can be hooked in, which makes it, at least theoretically, able to make or maybe even receive phone calls.
There’s a growing market in apps to improve or extend Auto-GPT and these are inter-connected by API key’s.
Like every other AI out there, Auto-GPT can make mistakes and although it can remember everything within a single session and even recover if the session is interrupted, it has no long-term memory.
But the “coolest thing” if we can describe things in those terms, about Auto-GPT is that it offers the promise of being able to do automation at the “mundane” level. For example, organizing a meeting – a complex task that most people find frustrating, could be easily automated and scheduled by an AI of this nature. And if that’s possible, what other use cases can be developed that even a small or medium sized business could adopt?
ITWC’s Research Lab is installing Auto-GPT and we’ll have more for you as we work through our cases. If you have any great ideas for use cases, we’ll see if we can work them into our research. Contact me directly at email@example.com
And coders and software engineers out there, if you are worried about losing your job to ChatGPT, we have some good news and some bad news.
Let’s start with the bad news (and get that out of the way). Another company has entered the code by AI race. Google announced last Friday that it’s AI software, Bard, can now be a “programming assistant.”
Bard can do code generation, debugging and explaining code snippets. It knows more than 20 program languages.
So with that, and Microsoft’s Co-Pilot for GitHub there is real competition out there.
Moreover, these programs are getting better and better. Where ChatGPT-3.5 could write a small program, version 4 is vastly improved and can take on much more complex and longer programs.
And for the frustrated QA analysts out there, both of these are probably better at commenting their code than the average programmer.
So what’s the good news? While this code is often solid, it is, according to some reports “not as secure as it should be.”
Four researchers from the Universite du Quebec in Canada have looked at the security of code generated by ChatGPT. They have published a paper with the title, “How Secure is Code Generated by ChatGPT?”
Spoiler alert – the answer, according to the computer scientists is “not very.”
“The results were worrisome,” the authors noted. “We found that, in several cases, the code generated by ChatGPT fell well below minimal security standards applicable in most contexts. In fact, when prodded to whether or not the produced code was secure, ChatGPT was able to recognize that it was not.”
Part of this might stem from safety features in ChatGPT itself. It’s designed to prevent it from writing code that can be used to attack other systems. Presumably, but we are not certain, Google’s Bard has the same restrictions.
But safety measures have been shown to make these models “less intelligent.”
Whatever the reason, code development needs to be carefully supervised by programmers who are able to spot flaws and as Google suggests, instruct the AI to address them, or put the coke and cheetos and, ugh, fix it manually.
The good news? It’ll do the commenting for you.
Sources include: University du Quebec researchers published at Cornell, The Register and CIODive
Sundar Pichai, CEO of Google 226 million dollars in compensation for 2022 may not be going over well with Google staff. Pichai makes approximately 800 times the median pay of the average Google employee.
This is at the same time as the company has cut about 12,000 jobs or 6 per cent of their global workforce.
Earlier this month, Google employees staged a walkout at the company’s London offices. In March employees walked out of the company’s Zurich offices after more than 200 workers were laid off.
Sources include: Reuters
It turns out that cybersecurity staff are not the only ones who are overwhelmed. Apparently business leaders are also suffering from decision anxiety, at least according to a study by Oracle in partnership with a firm called DKC Analytics.
The study covered 14,000 employees and business leaders globally, including 4,500 from six Asia-Pacific markets: Singapore, Australia, South Korea, China, India, and Japan.
74 percent of those surveyed in Asia-Pacific said the number of decisions that they had to make had increased “tenfold” in the last three years. Close to ninety percent said that it was making volume of data made decisions more complicated and was having a negative impact on their quality of life.
With the volume of data, a third of them said they didn’t know which sources of data to trust. And a shocking 71 per cent said they’d simply given up on making a decision.
In the wave of information overload, 92 per cent said they’d changed the way the made decisions. “96% said that they wanted to make decisions based on data but they felt that they “lacked the skills to interpret data in meaningful ways.”
So what are they doing? Close to a third said they relied on “their gut.” But 85 per cent would let a robot make their decisions.
So is this a phenomenon only for those in Asia and Australia? I checked with my robot decision maker and she said, “Sorry, I’m not sure about that.”
Great. So much for going to the Oracle.
Sources include: Oracle
Is Google losing its supremacy in providing support advice. We’ve all done it – you have a support issue and what do you do? We call, ask Dr. Google. You google it. And if someone comes to you for help without having Google’d it, they can expect to be heaped with disdain. That’s just how it works.
But there’s a new kid on the block. According to a new survey by a firm with the wonderful url of electric.ai which also has the great title “Who gives the best advice? ChatGPT vs Experts”
66 per cent of those who responded said they went to ChatGPT for help with their IT problems.
And in case you think it’s just GenZ who are doing this, the answer is – no – they are most likely to seek IT advice from ChatGPT with 83 per cent of them saying they’d go there first. But a big majority of Millennials would – at 67 per cent. Even half of GenX and (god forbid) Baby boomers would go to ChatGPT over Google.
So how accurate is ChatGPT? Electric asked 200 IT professionals to judge on the four most searched tech questions. ChatGPT had an overall accuracy score of 39 out of 100 – although some of the questions like, “How do I make my internet faster?” got a much larger accuracy score as high as 74 per cent.
Of course, if you live in the country like I do, the answer to “how do I make my internet faster is easy.”
Sources include: Electric.AI
That’s the top tech news for today. Hashtag Trending goes to air five days a week with the daily top tech news stories and we have a special weekend edition where we do an in depth interview with an expert on some tech development that is making the news.
Follow us on Apple, Google, Spotify or wherever you get your podcasts. Links to all the stories we’ve covered can be found in the text edition of this podcast at itworldcanada.com/podcasts
We love your comments. You can find me on LinkedIn, Twitter, or on Mastodon as @therealjimlove on our Mastodon site technews.social. Or just leave a comment under the text version at itworldcanada.com/podcasts
I’m your host, Jim Love, take deep breaths, relax and have a Marvelous Monday.
The post Hashtag Trending Apr.24th- Cybersecurity workers burnout; Code generated by ChatGPT and Google’s Bard not very secure; Execs would want a robot to make business decisions for them first appeared on IT World Canada.