DAI#42 – Scorching chips, unhealthy reminiscence, and AI ups and downs | DailyAI

Welcome to this week’s roundup of triple-distilled AI information.

This week, battle strains have been drawn within the chip wars in Taiwan.

Microsoft could must overlook to recollect.

And AI techs are combating for his or her proper to warn us about what they noticed.

Let’s dig in.

Made in Taiwan

This week all eyes have been on the Computex 2024 occasion in Taiwan, the place many of the world’s high-end AI chips are produced.

GPU prime canine NVIDIA unveiled its roadmap for more and more highly effective GPUs and different next-gen AI tech.

CEO Jensen Huang says the corporate shall be releasing a brand new household of GPUs yearly. The Blackwell GPUs are barely transport but, however Huang stated an upgraded model will launch later this yr.

If you take a nearer take a look at the Blackwell efficiency numbers, you start to get an thought of the potential affect NVIDIA’s new AI superchip will probably have.

AMD continues to be taking part in catchup within the AI chip race but in addition introduced that it’ll launch new generations of AI GPUs yearly. With its sights set on NVIDIA, AMD says its new AI chip, the Intuition MI325X GPU beats NVIDIA’s H200 flagship GPU.

The specs are significantly spectacular, nevertheless it will need to have been a bit of awkward bragging about beating the H200 when the Blackwell chips are virtually able to ship.

The AI chip battle between NVIDIA and just about each different computing firm ratcheted up a notch final week. Huge Tech firms, together with AMD and Intel, shaped a consortium to develop an open AI connectivity customary.

Guess who wasn’t invited to take part?

Neglect about it

The Silicon Valley mantra of ‘Move fast and break things’ was in proof this week.

Microsoft is about to unleash its Home windows AI Recall function which sounds nice on paper.

Recall remembers the whole lot you’ve seen and achieved in your PC and can assist out when you’ll be able to’t bear in mind on which web site you noticed that nice deal, or the place you saved your lolcat pics.

However there’s an issue. Tech safety specialists are saying the method Recall shops the information could possibly be a safety catastrophe. It’s worse than you suppose, and Microsoft could need to recall Recall.

Final week Google’s AI Overviews instructed customers to put glue on pizza and urged consuming rocks. Google lastly admitted AI Overviews has a couple of snags however says it’s largely your fault.

If Huge Tech can’t get AI search and working techniques proper, how will they deal with humanity’s large issues?

On the AI For Good Summit, business leaders mentioned examples of AI’s societal advantages and the challenges that must be overcome to keep away from widening the digital divide.

Will AI resolve a few of humanity’s largest issues or make them worse?

OpenAI’s ups and downs

What occurs after we develop into fully depending on AI instruments just for them to interrupt? This week we acquired a style of what that could possibly be like.

OpenAI’s ChatGPT had a significant outage as customers couldn’t entry the chat platform for hours. Was it a coincidence that Gemini, Claude, and Perplexity additionally had points across the identical time?

May it have something to do with supporters of disgruntled workers or AI skeptics?

Former and present OpenAI and Google workers simply printed a letter demanding the correct to warn about AI dangers with out penalties.

The “Right to Warn” letter makes some good factors however I’m undecided how probably it’s that OpenAI and others will decide to their calls for.

Pretend politics

Mendacity and politics go collectively like hearth and smoke. There by no means appears to be one with out the opposite.

AI has supercharged folks’s means to unfold disinformation however has additionally offered a helpful excuse for politicians in a decent spot.

Conservatives are demanding probably embarrassing audio of Joe Biden’s particular counsel interview. The Justice Division says it received’t launch the audio as a result of it could possibly be manipulated by AI.

Truthful sufficient, or grabbing at AI straws?

AI Occasions

In the event you’re seeking to be on the forefront of what’s taking place in AI then you definately’re spoilt for selection with occasions like these taking place world wide.

If you wish to see how AI can provide your advertising efforts a lift then take a look at the AI for B2B Entrepreneurs Summit 2024.

When you consider AI and electrical automobiles you instantly consider Tesla and self-driving. However AI additionally helps with thermal administration in EVs. The European EV Thermal Administration Summit 2024 explores the rising tendencies and tech on this area.

The AI_dev: Open Supply GenAI & ML Summit Europe is a dynamic two-day occasion specializing in open-source generative AI and machine studying. We’ve acquired an awesome low cost code for you in the event you’d wish to attend.

In different information…

Listed here are another clickworthy AI tales we loved this week:

And that’s a wrap.

On the fee new generations of GPUs are being launched, 2025 is ready to ship an enormous leap in AI advances. Are you able to think about trying again a yr from now and pondering how quaint GPT-4 is in comparison with GPT-5 and GPT-6?

Will you be utilizing Home windows AI Recall when it rolls out? I’d, however solely as soon as they cease storing the information in plain textual content in an unsecured folder. What have been they pondering?

The “Right to Warn” letter makes for chilling studying. When the individuals who constructed the AI say they’ve seen issues they’re not allowed to warn us about, we must always in all probability listen.

However Yann LeCun says there’s nothing to fret about so possibly it’ll be simply advantageous.

Is it time to panic, or time to cheer on the AGI and chill? Tell us what you suppose, and ship us hyperlinks to any attention-grabbing AI information tales we missed.

Recent articles