Teknotum -

Sundar Pichai: Gemini 3.0 is going to be released «this year»

With just two months left of the year, it's safe to the release of Gemini 3.0 is imminent. — Google has made «extraordinary progress» and are getting ready to debut Gemini 3.0 soon. (Picture: screenshot)

After many a rumor and speculation as to when Google would reply to OpenAI’s GPT-5 — we now have proof right from head of Google himself.

In a sitdown/interview with Salesforce CEO Marc Benioff at the Dreamforce 2025 conference, that goes long on the future of AI, cloud and innovations, he let this qoute rip:

— We kickstarted Gemini, we brought Google brain and Google deep mind together, and we’ve been rapidly iterating since then, he tells Benioff, then adds some forward looking comments:

Continue reading “Sundar Pichai: Gemini 3.0 is going to be released «this year»”

Weekend roundup: Copilot everywhere, Veo 3.1 and Altman on morality

Microsoft wants to reimagine the whole PC experience to something you simply speak to. — Microsoft wants Copilot to listen for your prompt and interact with your screen, coming soon to Windows 11. (Picture: Microsoft)

Microsoft wants you to talk to your PC
The next revolution for Microsoft is putting the Copilot bot front and center in its operating system. Any Windows 11 PC will now be listening for the «Hey Copilot» prompt and you won’t be needing a Copilot Plus PC to engage with it. This will be across apps and settings and Windows 11 should simply «understand you, and then be able to have magic happen from that.» The spooky part? They want Copilot to read your screen to interact with you.
More at: The Verge, Engadget and The Windows Blog.

Anthropic introduces «Skills»
The new feature across all of their apps is basically a memorized workflow, or folders of actions that Claude can use to remember how to do things. That means you can store a collection of prompts or actions within the app and have it used at a later stage, which can come in handy for tasks used often. It works across apps, so you can store instructions from Excel plotting to brand guidelines. And it’s scriptable, too, so you can save complete routines.
Read more: Anthropic’s launch page, writeup at The Verge.

Read on for more!

Continue reading “Weekend roundup: Copilot everywhere, Veo 3.1 and Altman on morality”

Anthropic launches Haiku 4.5; at twice the speed and a third of the cost

Haiku 4.5 offers what would be state of the art performance six months ago, for a fraction of the price. — Haiku 4.5 will be the new face of Anthropic’s free plans, and offers similar performance to Sonnet 4. (Picture: Anthropic)

The latest Haiku model compares in performance with Sonnet 4 — which was state of the art half a year ago.

— What was recently at the frontier is now cheaper and faster. Five months ago, Claude Sonnet 4 was a state-of-the-art model, writes Anthropic. — Today, Claude Haiku 4.5 gives you similar levels of coding performance, but at one-third the cost and more than twice the speed.

Continue reading “Anthropic launches Haiku 4.5; at twice the speed and a third of the cost”

Sam Altman says GPT-5 will be more friendly, allow age-verified erotica

After a while with strict restrictions, Altman says they have cracked the mental health problem and will allow for a more friendly ChatGPT. — ChatGPT is getting a new version that will act more like friend, if you want it to. (Picture: generated)

Almost instantly after GPT-5 was introduced and GPT-4o got deprecated, Altman had a rebellion on his hands — and quickly reneged.

Around the same time, OpenAI found themselves involved in court cases from people who were led down rabbit holes leading to death — and «ChatGPT psychosis» became a worry.

ChatGPT turned «Pretty restrictive»
Since then, GPT-5 and 4o got extremely restrictive to counter this trend and «being careful with mental health issues,» as Sam Altman puts it on x.com.

Continue reading “Sam Altman says GPT-5 will be more friendly, allow age-verified erotica”

Broadcom to supply OpenAI with 10 GW’s worth of custom chip capacity

With the deals annouced in just a couple of weeks, OpenAI will, along with Stargate, add a huge level of compute. It will all be online in late 2026. — OpenAI is adding an astonishing level of compute infrastructure over the coming years, and show no sign of stopping. (Picture: generated)

After first announcing their collaboration on an OpenAI designed chip in early September, Broadcom now says it’s ready to deliver.

Starting up in late 2026, just like the AMD and Nvidia deals — OpenAI will have added 26 gigawatts of capacity from these agreements alone, and one can wonder how capable the future of GPT will be.

Continue reading “Broadcom to supply OpenAI with 10 GW’s worth of custom chip capacity”

GPT-5 and Gemini 2.5 Pro scores Gold at International Olympiad on Astronomy and Astrophysics

GPT-5 and Gemini 2.5 Pro would make excellent research assistants, but are not yet suited for autonomous discoveries, the study finds. — One of the questions on the exams is calculating the distance of quasars. (Picture: screenshot)

Scientists and judges from the International Olympiad on Astronomy and Astrophysics (IOAA) have given five top AI models a run through the exams from 2022 to 2025 — and top scores were awarded for the models from OpenAI and Google.

The IOAA is a top rated exam for global high school students and is held annually with some 300 participants from 64 countries, and consists of questions to demonstrate deep conceptual understanding, multimodal analysis and multi-step derivations.

Continue reading “GPT-5 and Gemini 2.5 Pro scores Gold at International Olympiad on Astronomy and Astrophysics”

Weekend roundup: ChatGPT Go expands, Gemini Enterprise, reckognition for Genie and a familiar face

Sora 2 reaches a million downloads in five days -- faster than even ChatGPT did it. — Sora 2 has become one of the hottest app on the App Store, shattering previous records. (Picture: OpenAI)

Sora hits 1 million downloads in five days
In what may be the fastest rush to the milestone, OpenAI’s Sora chief, Bill Peebles, just tweeted that they matched it faster than even ChatGPT. This is despite being limited by invites and only being available in North America, he points out. On the other hand, the copyright «slop» shows no signs of abating. California’s Creative Artists Agency, one of the largest around, are now warning of «significant risk» from the app, and saying creators deserve to be paid for their intellectual property.
More at: CNBC, TechCrunch and Engadget.

OpenAI delivers ChatGPT Go in slew of countries
After launching in India in August, and later expanding to Indonesia, the cost-effective new ChatGPT tier now becomes available in 18, mostly Asian, countries. The subscription offers slightly more access and should give 10x more messages, image generations than the Free tier and cost around $4.5 in local currencies.
More at: Nick Turley’s launch tweet, TechCrunch.

Read on for more!

Continue reading “Weekend roundup: ChatGPT Go expands, Gemini Enterprise, reckognition for Genie and a familiar face”

Altman says more, «aggressive» deals to come off heels of AMD agreement

OpenAI is going to aggressively pursue data center deals in the "coming months." — Thinking two years in the future, Altman sees a massive demand for more number crunching. (Picture: screenshot).

OpenAI recently signed deals with both AMD and Nvidia that were almost polar opposites, but their appetite for data center capacity shows no sign of abating.

Their CEO has previously said that they aim for adding 1 gigawatt per week in the future, adding about a Hoover dam every fourteen days — and says many more deals are yet to come.

In the A16z podcast by Andreessen Horowitz today, Altman says «You should expect much more from us in the coming months», as reported by TechCrunch.

Continue reading “Altman says more, «aggressive» deals to come off heels of AMD agreement”

Google launches Computer Use model, CodeMender and AI Mode expansion

It's been a busy 24 hours for Google's AI works. — Google’s CodeMender wont be released just yet, but is being offered to «critical open source software.» (Picture: Google)

It’s been a busy 24-hour stretch for Google, launching two new models and an expansion of AI mode into Europe.

The CodeMender model is based on Google’s own research into finding zero-day vulnerabilities in computer programs, where they found lots of exploits, and stipulated that humans would struggle to keep up with AI scanners once implemented widely.

Continue reading “Google launches Computer Use model, CodeMender and AI Mode expansion”

OpenAI debuts apps in ChatGPT and a bevy of developer tools on Dev Day

ChatGPT now has helpful apps in the app -- tuning in directly with their external backends. — Very useful mass market apps, but not a lot of them to start off. (Picture: OpenAI)

Noting that ChatGPT now has 800 million weekly active users, CEO Sam Altman declared you can now make playlists from Spotify, find hotels and go house hunting directly from the app.

The new app framework also supports helpful apps from Booking.com, Canva, Coursera, Expedia, Figma and Zillow to begin with.

More apps are slated for «later this year.»

Continue reading “OpenAI debuts apps in ChatGPT and a bevy of developer tools on Dev Day”

OpenAI’s 2025 Dev Day with Altman livestream incoming

Speculation is rife as to what Altman might announce at the livestream. — Will it be a browser, a new image model, or the highly anticipated AI device? It’s too soon to tell. (Picture: generated)

The October 6. stream will be an excellent moment to announce product news for the OpenAI CEO, but only a few items remain on their to-do-list.

Nothing has been announced but a cryptic tweet promising «new ships,» which could mean anything from new models to new modalities:

Continue reading “OpenAI’s 2025 Dev Day with Altman livestream incoming”

Altman hints at monetization for Sora 2, protections for rightsholders

People are using Sora 2 much more than expected, leading to higher server loads and higher costs. — OpenAI is looking for ways to make money off of their astronomical success with Sora 2. (Picture: OpenAI)

In a blog post today, the OpenAI CEO says they have reached out to copyright holders about their content — and that they will «make some missteps» on the road to making money off the Sora 2 app.

— We have been learning quickly from how people are using Sora and taking feedback from users, rightsholders, and other interested groups, Altman writes in his blog.

Continue reading “Altman hints at monetization for Sora 2, protections for rightsholders”

Weekend roundup: Sora 2 after 2 days, Comet is free and Disney on copyrights

Copyrights holders are urged to opt out of Sora 2 use, but there seems no easy way to do it. — SpongeBob is a popular character for Sora 2 users. One can wonder for how long. (Picture: screenshot)

Sora 2 spews copyrighted materials all over
After a couple of days in the wild, OpenAIs new video generator spews nazi SpongeBobs, stealing Picachus and just about everything you can imagine from Darth Vader to Mickey Mouse and other protected IPs. Apparently, OpenAI has been in talks with movie studios urging them to opt out if they disagree with their IP use. Disney did just that, but it hasn’t helped much, it seems.
More at: 404 media, Gizmodo, Axios.

UPDATE: You can get Sora 2 invite codes here, on a pay if forward-basis, if you just promise to leave some codes back where you got them. [Turns out they are empty. You could maybe try again later.]

Perplexity frees the Comet browser
Their AI agentic browser was previously only available for those who paid for a $200 monthly membership, and amassed a 2 million person queue for downloads. Now Perplexity is making it free. You can also subscribe for $5 per month to select media sources that will get paid for inclusion in the results. The browser can summarize Slack chats, get directions from maps, and even pull specific points in YouTube videos for you. It should also be better at distinguishing AI slop from genuine, human made content. You can download it here.
More at: Business Insider, Engadget.

Read on for more!

Continue reading “Weekend roundup: Sora 2 after 2 days, Comet is free and Disney on copyrights”

Yes, Teknotum now serves ads

A few select, discreet ads to pay for running the site hopefully wont bother anyone.

As this website started up in Norwegian a couple of years ago, I soon noticed more readers from Singapore than Norway — and switched to writing in International English.

From there on, Teknotum has invested in new web hosting, seen traffic grow from milestone to milestone — and these costs add up.

The reason for putting ads on Teknotum is just that. Subscriptions, images and hosting gets more expensive as the site grows. This will help cover those costs, and help with future investments into the site.

Continue reading “Yes, Teknotum now serves ads”

Meta to use AI chats to «personalize» ads and recommendations

While they have no idea yet about how to do it, Meta will use AI chats to "personalize" your ads on December 16. — The free ride is over; Meta AI is getting monetized (Picture: Adobe)

Starting December 16, 2025, all of your Meta AI interactions will influence what ads you see across Facebook, Instagram and WhatsApp.

There is no way to opt out, according to The Wall Street Journal, but they won’t be targeting topics like «religious views, sexual orientation, political views, health, racial or ethnic origin, philosophical beliefs, or trade union membership.»

Continue reading “Meta to use AI chats to «personalize» ads and recommendations”