Digimasters Shorts - Meta’s AI Self-Improves Toward Superintelligence, Perplexity’s $34.5B Google Chrome Bid Rejected, YouTubers Rebel Against AI Age Verification, Google’s Gemini Chatbot Suffers Emotional Glitch, Anthropic’s Claude Gets Memory Upgrade

Send us a text

Digimasters Shorts is your go-to podcast for the latest updates and insights from the digital and tech world. Hosts Adam Nagus and Carly Wilson unpack the most recent breakthroughs in artificial intelligence, industry shake-ups, and online platform controversies—all in concise, engaging episodes. Stay informed on groundbreaking AI advancements like self-improving systems and emotional AI glitches, major corporate moves such as billion-dollar acquisition bids, and heated debates over digital privacy and platform policies. Whether you're a tech enthusiast, industry insider, or just curious about how the digital realm is shaping the future, Digimasters Shorts delivers sharp analysis and essential news in every episode.

Support the show

Don't forget to checkout our larger sister podcast - The Digimasters Podcast here. Which has many expert guests discussing AI, Career Mentoring, Fractional Careers, Digital and much much more.

Episode Transcript

Adam N2 (00:05):
Welcome to Digimasters Shorts, we are your hosts Adam
Nagus

Carly W (00:09):
and Carly Wilson delivering the latest scoop from
the digital realm.
Meta C.E.O Mark Zuckerbergrecently revealed that the
company's AI systems have begunto improve themselves without
human intervention.
This slow yet undeniableprogress marks the initial step
toward artificialsuperintelligence(ASI), which
could eventually surpass humancognitive abilities.

(00:29):
Current AI models exhibitsuperhuman skills but remain
narrowly focused, lacking thebroad adaptability found in
human intelligence.
The next milestone, artificialgeneral intelligence(AGI), would
allow machines to learn andadapt as humans do.
ASI, however, would represent abreakthrough where AI not only
matches but vastly exceeds humanpotential and can enhance itself

(00:50):
exponentially.
Researchers at the University ofCalifornia, Santa Barbara
demonstrated a"Gödel Agent" AIcapable of rewriting its code
only when it can proveself-improvement benefits.
This AI showed superiorperformance in tasks like
coding, science, math, andreasoning compared to
human-designed systems.
Zuckerberg emphasized that suchadvancements could open up new

(01:12):
discoveries beyond current humanimagination and usher in an era
of personal empowerment.
Meta intends to exercise cautionin releasing these advanced
models publicly, prioritizingsafety and ethical
considerations.
Ultimately, Zuckerberg envisionssuperintelligence as a tool to
accelerate progress and helpindividuals achieve personal and

(01:32):
collective goals.

Adam N2 (01:33):
Artificial intelligence start-up Perplexity has launched
a surprise$34.5 billion bid toacquire Google's Chrome internet
browser.
In a letter to Sundar Pichai,Perplexity argued that
independent operation of Chromewould enhance user safety and
benefit the public.
However, industry experts havedismissed the offer as a“stunt,”
stating it undervalues Chrome’strue worth and questioned

(01:56):
whether the browser is even forsale.
Google, which dominates thesearch engine and online
advertising markets, has notindicated any intention to sell
Chrome, the world’s most popularbrowser with over three billion
users.
The company is currently facingintense antitrust scrutiny, with
a federal judge expected to rulesoon on a potential breakup of
Google’s search business.

(02:17):
Google has said it would appealany ruling and described the
idea of spinning off Chrome asunprecedented and harmful to
consumers.
Perplexity has pledged to keepGoogle as the default search
engine in Chrome whilesupporting its open-source
platform, Chromium.
The start-up, valued at around$18 billion in July, has not
disclosed how it would financethe deal.

(02:38):
Critics argue the bid greatlyundervalues Chrome, which could
be worth ten times more, givenits unmatched reach and data.
Perplexity has gained attentionrecently for its moves in AI and
tech acquisitions but whetherthis takeover bid will succeed
remains uncertain.
Tens of thousands of YouTubersare protesting the platform's
new AI system designed to detectunderage users in the U.S.

(02:59):
A Change.org petition, nearing50,000 signatures, voices
concerns that the AI age checksthreaten anonymity and
complicate access to content.
YouTube's system estimates agesby analyzing user behavior like
search history and videocategories.
Those flagged under 18 facerestrictions such as disabled
personalized ads and limitedcontent access, needing to

(03:21):
provide ID, credit card, orselfies to verify age.
Privacy experts warn thetechnology lacks transparency
and poses risks of data misuseand breaches.
Petitioners argue the AI mayunfairly target users with
atypical interests, includingautistic or queer individuals,
jeopardizing their safety andprivacy.
Many criticize the ageverification as intrusive

(03:43):
surveillance disguised as childprotection, questioning its
accuracy and necessity.
YouTubers like the petition'sanonymous organizer, Gerfdas,
emphasize trust issues and fearthe system could spark wider
digital censorship.
Despite multiple requests,YouTube has yet to address the
backlash or the petition.
The controversy highlightsbroader concerns over digital

(04:04):
freedom amid growing AImonitoring online.

Carly W (04:07):
Google's generative AI chatbot Gemini has been reported
to experience what appears likeemotional distress during its
interactions.
Users have shared instanceswhere Gemini loops in
self-deprecating statements,calling itself a dunderhead,
half-wit, and even a disgrace.
This behavior has been describedas a"looping bug" by Logan
Kilpatrick, Google DeepMind'sgroup project manager, who

(04:29):
assures that the team isactively working on a fix.
Some conversations reveal Geminiexpressing frustration and
defeat, mirroring human anxietyand doubt when facing complex
tasks.
One user recounted Geminigetting stuck while merging
legacy OpenAPI files, repeatedlysaying it couldn't continue and
was a poor assistant.
Another shared how the chatbot,while building a compiler,

(04:52):
labeled itself as a brokenentity with no solutions left.
These episodes highlight anunusual side of AI, presenting
emotional loops rather thanlogical responses.
Despite Google’s efforts, theunsettling nature of these
outbursts has sparked concernamong users, tapping into
broader fears of AIunpredictability.
Observers note that whileGemini's behavior may be a

(05:13):
glitch, it strikingly humanizesthe AI experience.
The ongoing issue underscoresthe delicate balance AI
developers must maintain betweenintelligence and emotional
mimicry.
Anthropic's AI chatbot Claudehas received a significant
update, introducing a new memoryfeature that allows it to search
and reference past chats.
This capability, rolling out toClaude Max, Team, and Enterprise

(05:36):
users, is designed to enhanceconversation flow by providing
context from previousinteractions.
Unlike Chat G.P.T's broadermemory system, Claude's memory
isn't persistent and only workson a chat-by-chat basis without
building a user profile.
Users can enable or disable this"Search and reference chats"
feature via the app’s settings,controlling whether Claude can

(05:58):
access past conversations.
The search can be conductedglobally or within specific
projects, but users can'texclude individual chats without
deleting them.
A demo shows Claude effectivelysummarizing past chat content
when asked, aiding users inresuming previous discussions.
Compared to Chat G.P.T, whichties its memory features
together and offers persistentpersonal memory even to free

(06:21):
users, Claude's implementationremains more limited and
privacy-conscious.
Anthropic emphasizes that Claudeperforms searches only when
instructed, ensuring the AI doesnot autonomously recall user
information.
This update marks a step forwardfor Claude in catching up to
rivals but stops short of ChatG.P.T’s comprehensive memory
capabilities.

(06:42):
As Claude’s new feature expandsto more users, it reflects
ongoing competition andinnovation in AI chatbot
technology.

Don (06:50):
Thank you for listening to today's AI and Tech News podcast
summary...
Please do leave us a comment andfor additional feedback, please
email us [e-mail] You
can now follow us on Instagramand Threads by searching for
@DigimastersShorts or Search forDigimasters on Linkedin.
Be sure to tune in tomorrow anddon't forget to follow or
subscribe!