Well, I’m back – what did I miss?
It seems like we’re discovering the real costs of AI – both in training and other impacts. Deep fakes on Tik Tok let students terrorize middle school teachers? And more layoffs in the Bay Area – is this the sign of more to come elsewhere?
All this and more on the “a fool knows the cost of everything and value of nothing” edition of Hashtag Trending. I’m your host Jim Love, let’s get into it.
“A new wave of AI hype is making bold claims about the future of knowledge production. Mustafa Suleyman, DeepMind co-founder and Microsoft AI CEO, recently predicted at the Aspen Ideas Festival that AI will reduce the cost of producing knowledge to zero.
Suleyman stated, ‘In 15 or 20 years’ time, we will be producing new scientific, cultural knowledge at almost zero marginal cost. It will be widely open sourced and available to everybody.’
However, this vision faces several challenges. The upfront costs of AI development are astronomical, with the tech industry pouring hundreds of billions into AI infrastructure. Moreover, knowledge is more than just information – it requires human learning and experience.
Critics argue that AI can distribute existing knowledge but may not truly ‘produce’ new, original insights. As one expert points out, ‘AIs can’t do field research, design new experiments or create great artworks that respond to the feeling of being alive in a particular moment.’
While AI has potential in scientific research and creative processes, the idea of cost-free knowledge production may be overly optimistic. As we’ve seen with past tech booms, the reality often falls short of the hype. The future of AI and knowledge production is likely to be more nuanced, with both costs and benefits to consider.”
Sources include: Axios
And speaking of costs, Anthropic’s CEO Daio Amodei revealed in a recent podcast that while current models like ChatGPT-4 cost about $100 million to train, some models in development today are approaching the $1 billion mark.
Anthropic has just done a big release of its Claude 3 model which many believe is equivalent or in some aspects superior to OpenAI’s 4o release. Claude has achieved real prominence in terms of its ability as a writing platform and is getting rave reviews in terms of software development. But the costs are significant.
Amodei predicts even more staggering figures in the near future. He states, ‘I think if we go to ten or a hundred billion, and I think that will happen in 2025, 2026, maybe 2027… then I think there is a good chance that by that time we’ll be able to get models that are better than most humans at most things.’
This exponential growth in AI capabilities comes with equally exponential hardware demands. Last year alone, over 3.8 million GPUs were delivered to data centers. With Nvidia’s latest AI chip costing up to $40,000, the numbers add up quickly.
But it’s not just about chips. Power consumption is a growing concern. The GPUs sold to data centers last year could power 1.3 million homes. This surge in energy needs is prompting tech giants like Microsoft to explore alternative power sources, including modular nuclear reactors.
As AI continues its rapid advance, the question remains: how will these increasingly powerful and costly models shape our future?”
Sources include: Tom’s Hardware
Speaking of the cost of power consumption –
“Google’s latest environmental report reveals a complex relationship between AI and climate change. While the tech giant promotes AI as a tool to combat global warming, it’s also driving up the company’s own carbon footprint.
Google’s corporate emissions rose 13% last year and are up 48% since 2019. A key factor in this increase is the growing power demand from data centers serving AI and other applications.
In the report, Google acknowledges the ‘challenge of reducing emissions while compute intensity increases and we grow our technical infrastructure investment to support this AI transition.’
However, the company isn’t giving up on its environmental goals. Google says they are actively working on making AI infrastructure more efficient and highlights how its AI products are helping reduce emissions in other sectors. For example, they’ve developed tools that cities can use to improve traffic flow and reduce vehicle emissions.
But Google is facing a steep climb to reach its ambitious 2030 net-zero goal. The report underscores the complex balance between advancing AI technology and managing its environmental impact, a challenge that will likely define the tech industry’s relationship with climate change in the coming years.”
And here’s a story that knocked me for a loop:
“A disturbing trend has emerged at Great Valley Middle School in Pennsylvania, where students launched a coordinated attack on teachers using TikTok. Over 20 educators discovered fake accounts impersonating them, posting disparaging, lewd, and offensive content.
Patrice Motz, a Spanish teacher, found a fake profile using her family photo with a disturbing caption implying pedophilia. Social studies teacher Shawn Whitelock said, ‘An impersonator assassinated my character – and slandered me and my family in the process.’
The incident has left many teachers feeling violated and demoralized. Bettina Scibilia, an English teacher, noted, ‘Many of my students spend hours on TikTok, and I think it’s just desensitized them to the fact that we’re real people.’
While the school has taken some disciplinary action, legal limitations on regulating off-campus speech have hampered their response. The local teachers’ union has called for stronger measures to prevent such behavior.
This attack represents a significant escalation in student-teacher conflicts on social media, raising concerns about the impact of technology on empathy and respect in schools.”
Sources include: New York Times
“The Bay Area tech industry is facing its worst wave of layoffs in over a year, signaling that the sector’s job cuts are far from over. During the second quarter of 2024, tech companies announced plans to eliminate more than 7,000 jobs in the region.
This marks the highest quarterly total since early 2023, when over 10,000 tech jobs were cut. The trend has continued into recent weeks, with companies like Lacework, Moxion Power, and Planet Labs announcing hundreds of layoffs in late June alone.
Since 2022, tech firms have revealed plans to cut nearly 45,000 jobs in the Bay Area. This includes about 10,300 in 2022, nearly 21,600 in 2023, and approximately 13,000 in just the first half of 2024.
These layoffs span various tech sectors, from cloud security and green energy to satellite technology and telecommunications. Companies cite efficiency drives and efforts to improve their bottom lines as reasons for the cuts.
The persistent layoffs raise concerns about the tech industry’s stability and its impact on the Bay Area’s economy, traditionally reliant on this sector for job growth and innovation.”
Sources include: TechXplore.com
And that’s our show for today.
Hashtag Trending is on summer hours. We will have 3 daily news shows a week, with a weekend interview show we call the Weekend Edition. And we’ll hopefully have one more week off in the summer before we come back to a full schedule on Labour Day.
Show notes are at technewsday.ca or .com – either one works.
We love your comments. Contact me at editorial@technewsday.ca
I’m your host Jim Love, have a Marvelous Monday.
