Federal regulators launched an investigation into Tesla's self-driving robotaxi service just one day after its celebration, citing multiple safety violations captured on video. Meanwhile, Intel announced the shutdown of its automotive architecture business, continuing a series of layoffs and restructurings under new CEO Lib-Bu Tan. In AI news, Google co-founder Sergi Brin's suggestion to threaten AI for better performance has been debunked, with experts recommending context and clarity instead of aggression for improved AI interactions. Host Jim Love discusses these developments and their implications in the latest episode of Hashtag Trending.
00:00 Tesla's Robotaxi Launch Faces Immediate Scrutiny
02:46 Intel Shuts Down Automotive Business Amid Layoffs
05:21 Google Co-founder's Controversial AI Advice
07:13 The Power of Politeness in AI Interactions
08:33 Conclusion and Contact Information
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Federal regulators may spoilTesla's celebration of the robotaxi.
launch.
Intel shuts down its automotivebusiness as layoffs continue and
Google's co-founder's advice tothreaten AI won't work anymore.
Welcome to Hashtag Trending.
I'm your host, Jim Love.
Let's get into it.
Tesla declared its robotaxi launched theculmination of a decade of hard work,
(00:24):
but federal safety regulators launched aninvestigation within 24 hours after video
showed the company's self-driving carsviolating traffic laws on Austin roads.
The contrast between Tesla'svictory lap and regulatory reality
tells the story while Elon Muskcelebrated on social media.
Federal investigators reviewedfootage of robo taxis driving
(00:47):
into oncoming traffic lanes.
Bloomberg reported that US Auto Safetyregulators are looking into incidents
where Tesla Inc's self-drivingrobotaxis appeared to violate traffic
laws during the company's first dayoffering of paid rides in Austin.
a national highway and TrafficSafety Association spokesperson
(01:08):
Confirmed that the agency is awareof the referenced incidents and is
in contact with the manufacturerto gather additional information.
There have been numerous reports ofissues from the Trials Tech Journal.
The Verge is maintaining what they'recalling a running list of robo Taxii
mishaps from just the first three days,including driving on the wrong side of
(01:29):
the street, phantom braking incidents,dropping passengers in busy intersections,
failing to recognize reversing trucksand driving over curbs, All from
a fleet of just 10 to 20 vehiclesin heavily controlled conditions.
A YouTube host who featured a Teslarobotaxi competition with Waymo not
only lost the supposed race with theWaymo vehicle, but was also hit with
(01:53):
phantom problems during her videoA popular automotive news website,
Jalopnik noted the absurdity, Tesla'srobotaxis faced federal investigations
with just one day on the road.
Among the many items that Jalopnikreported was the robotaxi cutting across
double yellow lines into oncoming traffic.
(02:14):
Tesla had earlier tried to controlinformation by trying to block
Austin from releasing publicrecords about the trial to Reuters.
It seems like Reuters wasthe least of their problems.
This trial was bound to facescrutiny from a vast number of
publications and citizen journalists,but it does raise the question.
If 10 to 20 vehicles produce thesedocumented problems in just three days,
(02:38):
Musk's plan for thousands within monthsraises some serious safety concerns.
Intel's new CEO continued the rapidrestructuring of the troubled giant with
the announcement that they were shutteringmost of Intel's automotive architecture
business and laying off most of the staff
marking another step in thecompany's retreat from what they're
(02:59):
calling non-core operations.
According to media reports, Intelstated, we've decided to wind down the
automotive business within our clientcomputing group after an internal memo
was shared with employees Tuesday morning.
The move comes just months afterthe division showcased AI enhanced
vehicle technology at CES 2025and met with Chinese automakers at
(03:23):
the Shanghai Auto Show in April.
Intel spokesperson Cory PforzheimerHeimer said, as we have said
previously, we were refocusing ourcore client and data center portfolio
to strengthen our product offerings.
The automotive shutdown is part ofa massive restructuring under new
CEO Lib-Bu Tan, who plans to lay offmore than 20% of Intel's workforce,
(03:45):
potentially over 21,000 employees.
Intel already cut 15,000 jobs last Augustand filed Warren notices for 107 more
California layoffs starting July 15th.
We reported yesterday on thecompany's decision to slash
employment and its marketingfunctions and work with Accenture
to replace those employees with ai
(04:08):
In the automotive sector, Intel maynot have as large a footprint as its
competitors, but the company claims 50million vehicles use its processors.
Though it doesn't break out theautomotive revenue or headcount, the
company promises that it will finishthe projects that it has underway and
The shutdown won't affect Intel'sseparate investment in Mobile Eye,
(04:29):
the Israeli autonomous drivingtechnology company where Intel holds a
controlling stake Mobile Eye developsself-driving car software and vision
systems that power automated drivingfeatures in vehicles worldwide,
operating independently from Intel'snow closed automotive chip division.
Intel is targeting $1.5 billion inoperating expense cuts over two years
(04:53):
while spinning off non-core divisions.
The company has lost ground tocompetitors like NVIDIA in AI chips
and TSMC in manufacturing with itsstock dropping 67% over five years.
And as we predicted in our last report,watch for even more Kats as CEO.
Tan attempts to stem the huge lossesand find some areas where the company
(05:17):
can reestablish its market leadership.
Google Co-founder Sergi Brin recentlyclaimed that AI models work better
if you threaten them with physicalviolence, suggesting that threatening
to kidnap a chatbot improves responses.
People feel weird about it,so we don't talk about it.
He said, and he was right about that.
(05:38):
It feels weird.
I mean, how can you be mean to an AIthat always talks so nicely to you?
That's not what nice people do, and itturns out there's a lot of nice people.
People like to be polite.
So much so in fact, that ChatGPT CEO,Sam Altman had remarked that OpenAI
was paying millions of dollars tosupport these polite interactions.
(06:01):
Apparently even please and thank youor have a nice day or, that's terrific.
All of that has to be analyzed andprocessed, which takes extra computing
cycles, So it's a bit of a victory forthe nice and polite, the people that
feel that Brin's tip goes against ournicer nature that apparently it doesn't
work anymore, and testing has proven it.
(06:24):
Chris Hoffman, who's known for theWindows Intelligence Newsletter, decided
to test Brin's theory that threatsmade generative AI perform better
and got immediate proof it's wrong.
In fact, when he tried threateningChatGPT, he didn't even get a response.
Just content removed with a warningabout violating usage policies.
(06:44):
Brin's claim may have worked withan earlier AI version where high
stakes language, like if you don'tget this right, I'll lose my job.
Seem to improve responses.
But apparently AI hasgrown beyond that now.
as Hoffman's tests showed, modern chatbotslike ChatGPT, either have safety filters
that block threatening language, orthey've learned that they don't have
(07:05):
to take this sort of abuse, whatever.
It makes the approach counterproductive.
But Hoffman had an idea.
What if it wasn't the threat?
What if it was the signal for urgency thatwould make AI more careful and precise?
Because language models predict textpatterns and urgent contexts are
associated with higher quality responses.
(07:28):
And
As it turns out, there may be ahope for those of us, especially the
Canadians in our audience who demand ourright to be polite and compassionate.
So instead of threats, Hoffman recommendsproviding context and clarity rather than
give me a driving route or I'll kidnap you
I need a route with stops every two hours'cause my poor dog needs regular breaks.
(07:50):
The specific context appears to producebetter and more relevant results.
Once users could explain whysomething matters or provide helpful
details about what they want,the context improves the answers.
Politeness may be optional, but insofar asit leads to straightforward communication
and greater context, it might be worththe money that Sam Altman claims it costs.
(08:14):
So instead of a threat, either give itmore context or just try to be friendly.
This really matters.
Think hard about this.
Okay.
Um, you can follow that by saying,please, it might not help, but as
the saying goes, it can't hide.
And that's our show.
(08:35):
We'd love to hear from you.
Even those of you whoare a little bit cranky.
You can reach us at our new improvedwebsite@technewsday.ca or.com.
take your pick.
Just go to the Contact us tab andyou can also find past show notes,
podcasts, and some news there as well.
I'm your host, Jim Love.
Have a thrilling Thursday please.
Federal regulators may spoilTesla's celebration of the robotaxi.
launch.
Intel shuts down its automotivebusiness as layoffs continue and
Google's co-founder's advice tothreaten AI won't work anymore.
Welcome to Hashtag Trending.
I'm your host, Jim Love.
Let's get into it.
Tesla declared its robotaxi launched theculmination of a decade of hard work,
(00:24):
but federal safety regulators launched aninvestigation within 24 hours after video
showed the company's self-driving carsviolating traffic laws on Austin roads.
The contrast between Tesla'svictory lap and regulatory reality
tells the story while Elon Muskcelebrated on social media.
Federal investigators reviewedfootage of robo taxis driving
(00:47):
into oncoming traffic lanes.
Bloomberg reported that US Auto Safetyregulators are looking into incidents
where Tesla Inc's self-drivingrobotaxis appeared to violate traffic
laws during the company's first dayoffering of paid rides in Austin.
a national highway and TrafficSafety Association spokesperson
(01:08):
Confirmed that the agency is awareof the referenced incidents and is
in contact with the manufacturerto gather additional information.
There have been numerous reports ofissues from the Trials Tech Journal.
The Verge is maintaining what they'recalling a running list of robo Taxii
mishaps from just the first three days,including driving on the wrong side of
(01:29):
the street, phantom braking incidents,dropping passengers in busy intersections,
failing to recognize reversing trucksand driving over curbs, All from
a fleet of just 10 to 20 vehiclesin heavily controlled conditions.
A YouTube host who featured a Teslarobotaxi competition with Waymo not
only lost the supposed race with theWaymo vehicle, but was also hit with
(01:53):
phantom problems during her videoA popular automotive news website,
Jalopnik noted the absurdity, Tesla'srobotaxis faced federal investigations
with just one day on the road.
Among the many items that Jalopnikreported was the robotaxi cutting across
double yellow lines into oncoming traffic.
(02:14):
Tesla had earlier tried to controlinformation by trying to block
Austin from releasing publicrecords about the trial to Reuters.
It seems like Reuters wasthe least of their problems.
This trial was bound to facescrutiny from a vast number of
publications and citizen journalists,but it does raise the question.
If 10 to 20 vehicles produce thesedocumented problems in just three days,
(02:38):
Musk's plan for thousands within monthsraises some serious safety concerns.
Intel's new CEO continued the rapidrestructuring of the troubled giant with
the announcement that they were shutteringmost of Intel's automotive architecture
business and laying off most of the staff
marking another step in thecompany's retreat from what they're
(02:59):
calling non-core operations.
According to media reports, Intelstated, we've decided to wind down the
automotive business within our clientcomputing group after an internal memo
was shared with employees Tuesday morning.
The move comes just months afterthe division showcased AI enhanced
vehicle technology at CES 2025and met with Chinese automakers at
(03:23):
the Shanghai Auto Show in April.
Intel spokesperson Cory PforzheimerHeimer said, as we have said
previously, we were refocusing ourcore client and data center portfolio
to strengthen our product offerings.
The automotive shutdown is part ofa massive restructuring under new
CEO Lib-Bu Tan, who plans to lay offmore than 20% of Intel's workforce,
(03:45):
potentially over 21,000 employees.
Intel already cut 15,000 jobs last Augustand filed Warren notices for 107 more
California layoffs starting July 15th.
We reported yesterday on thecompany's decision to slash
employment and its marketingfunctions and work with Accenture
to replace those employees with ai
(04:08):
In the automotive sector, Intel maynot have as large a footprint as its
competitors, but the company claims 50million vehicles use its processors.
Though it doesn't break out theautomotive revenue or headcount, the
company promises that it will finishthe projects that it has underway and
The shutdown won't affect Intel'sseparate investment in Mobile Eye,
(04:29):
the Israeli autonomous drivingtechnology company where Intel holds a
controlling stake Mobile Eye developsself-driving car software and vision
systems that power automated drivingfeatures in vehicles worldwide,
operating independently from Intel'snow closed automotive chip division.
Intel is targeting $1.5 billion inoperating expense cuts over two years
(04:53):
while spinning off non-core divisions.
The company has lost ground tocompetitors like NVIDIA in AI chips
and TSMC in manufacturing with itsstock dropping 67% over five years.
And as we predicted in our last report,watch for even more Kats as CEO.
Tan attempts to stem the huge lossesand find some areas where the company
(05:17):
can reestablish its market leadership.
Google Co-founder Sergi Brin recentlyclaimed that AI models work better
if you threaten them with physicalviolence, suggesting that threatening
to kidnap a chatbot improves responses.
People feel weird about it,so we don't talk about it.
He said, and he was right about that.
(05:38):
It feels weird.
I mean, how can you be mean to an AIthat always talks so nicely to you?
That's not what nice people do, and itturns out there's a lot of nice people.
People like to be polite.
So much so in fact, that ChatGPT CEO,Sam Altman had remarked that OpenAI
was paying millions of dollars tosupport these polite interactions.
(06:01):
Apparently even please and thank youor have a nice day or, that's terrific.
All of that has to be analyzed andprocessed, which takes extra computing
cycles, So it's a bit of a victory forthe nice and polite, the people that
feel that Brin's tip goes against ournicer nature that apparently it doesn't
work anymore, and testing has proven it.
(06:24):
Chris Hoffman, who's known for theWindows Intelligence Newsletter, decided
to test Brin's theory that threatsmade generative AI perform better
and got immediate proof it's wrong.
In fact, when he tried threateningChatGPT, he didn't even get a response.
Just content removed with a warningabout violating usage policies.
(06:44):
Brin's claim may have worked withan earlier AI version where high
stakes language, like if you don'tget this right, I'll lose my job.
Seem to improve responses.
But apparently AI hasgrown beyond that now.
as Hoffman's tests showed, modern chatbotslike ChatGPT, either have safety filters
that block threatening language, orthey've learned that they don't have
(07:05):
to take this sort of abuse, whatever.
It makes the approach counterproductive.
But Hoffman had an idea.
What if it wasn't the threat?
What if it was the signal for urgency thatwould make AI more careful and precise?
Because language models predict textpatterns and urgent contexts are
associated with higher quality responses.
(07:28):
And
As it turns out, there may be ahope for those of us, especially the
Canadians in our audience who demand ourright to be polite and compassionate.
So instead of threats, Hoffman recommendsproviding context and clarity rather than
give me a driving route or I'll kidnap you
I need a route with stops every two hours'cause my poor dog needs regular breaks.
(07:50):
The specific context appears to producebetter and more relevant results.
Once users could explain whysomething matters or provide helpful
details about what they want,the context improves the answers.
Politeness may be optional, but insofar asit leads to straightforward communication
and greater context, it might be worththe money that Sam Altman claims it costs.
(08:14):
So instead of a threat, either give itmore context or just try to be friendly.
This really matters.
Think hard about this.
Okay.
Um, you can follow that by saying,please, it might not help, but as
the saying goes, it can't hide.
And that's our show.
(08:35):
We'd love to hear from you.
Even those of you whoare a little bit cranky.
You can reach us at our new improvedwebsite@technewsday.ca or.com.
take your pick.
Just go to the Contact us tab andyou can also find past show notes,
podcasts, and some news there as well.
I'm your host, Jim Love.
Have a thrilling Thursday please.
Popular Podcasts
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.