• The ban, if implemented, would mean that any printer without this firearm-blocking technology could not legally be sold or transferred in California
• The move is still pending Senate review before being enacted into law
• It could mean that 3D printers could become more expensive or more restrictive for users in the state in the near future as manufacturers pass on added compliance costs to users

The state of California is moving toward enacting a ban on the sale of 3D printers that lack a built-in algorithm preventing users from producing ‘ghost guns’ on a whim.

The controversial bill was passed last week and is pending Senate confirmation before ultimately reaching California Governor Gavin Newsom’s desk, where it must still be signed.

The state law governing app use by minors takes effect tomorrow.

• The upcoming Phison E37T SSD controller manages to max out PCI-E 5.0 read speed ratings at 14,900 MB/s
• It consumes less than a 3rd of the power of Phison’s older DRAM-infused SSD controllers
• The move is a crucial one, aligned with Phison’s anticipation of high DRAM prices in the foreseeable future, aimed at providing relief to mainstream and enthusiast consumers

With SSD prices moving sharply upwards over the last few months, thanks to unrelenting AI demand across the board, consumers are increasingly looking to the lower end of the spectrum to bridge the gap between their budgets and the cost of modern SSDs.

The upcoming Phison E37T SSD controller could help tide things over. It happens to be the first Gen 5 DRAM-less SSD controller to max out bandwidth over the 4 lanes available to an M2 SSD on modern PCs, laptops, and consoles.

In the world of buzzwords, the acronym ‘AI’ has absolutely been the buzziest of buzzing buzzwords for at least a few years now. Where previously terms like ‘smart’ and ‘intelligent’ sufficed to promote a product, we are now being told that we are living in an age where this supposedly newfangled ‘artificial intelligence’ is doing literally everything faster and better while also curing cancer on the side. Yet, as a wise man once said: “You keep using that word. I do not think it means what you think it means.”

The obvious implication of using a term like ‘artificial intelligence’ in this manner is that it brings to mind a modern version of early last century’s ‘electronic brain’ vernacular alongside the rise of digital computers. Yet rather than electrons in vacuum tubes and semiconductors propelling us into a brave new world of super-intelligence, we now just use said devices to doom scroll and to engage in passive-aggressive online communications like the typical primate groups in a virtual jungle defending their turf.

Similarly, the term AI is massively oversold today, least of all in the inherent presupposition that we somehow have finally cracked the mystery of the brain and have created an intelligence that can go toe-to-toe with humans and even our corvid dinosaur friends. Perhaps the worst part is that there is a veritable mountain of fascinating algorithms and other constructs that help us automate many tasks today, making it somewhat rude to just give up and call everything ‘AI’ like we learned nothing from the 1980s AI craze.

So what is exactly being smoothed over by the glossy marketing of ‘everything is AI’?

Cognition Versus Intelligence

Recently I covered the topic of intelligence, both in the sense of its definition and the empirical evidence. Within that definition it is already quite obvious that animals like birds are pretty intelligent, and can compete with the average human on a number of metrics. Of the different types of intelligence, fluid intelligence (Gf) is perhaps the most crucial since it pertains to what might be the clearest sign of intelligence in the form of reasoning.

Add to this memory (knowledge and recall) as well as acquired skills and you got the basics of general intelligence. One could absolutely make the point that this is all that intelligence is about, as in the acquisition of data, processing it and using reasoning to come to new conclusions. Yet as can be seen in the referenced article, the basic CHC intelligence model can, and has been, expanded to include sensory, motor and efficiency metrics, which are very species-centric.

Of course, it is true that within cognitive processes it’s hard to exclude sensory input and output via actuators like muscles to perform some kind of physical action. After all, no type of intelligence is of much use if there are no in- and output, such as how we need at least one of our five senses to be aware of the world around us along with some way to interact. Whether intelligence could develop without both is also a valid question.

The resulting disagreements in the academic community on where to draw the line between intelligence and cognition do not help with narrowing the scope of ‘intelligence’, as it makes it possible to assign the label to something like machine vision. Even when this is a system that merely replicates parts of the visual cognitive process without the underlying reasoning and understanding that accompanies this cognitive process in us animals.

What we can conclude from this, however, is that what we call ‘smart’ or ‘AI’ are merely systems that attempt to replicate such a fragment of the human cognitive process.

Machine Vision

Perhaps the biggest strength of machine vision (MV) is that it allows for a cognitive task to be off-loaded to a computer system that will never suffer fatigue or become distracted. This is essential in tasks like quality assurance, such as on production lines. Rather than having a human check each item that zips past for certain properties, alignments, etc. a machine vision system can take over this cognitive task while being inarguably far more efficient.

MV encompasses a wide range of implementations, all targeting a specific task that can use different sensors and outputs to accomplish a goal. For e.g. PCB assembly lines and food production you got many MV systems that use visible light as well as near-infrared and other camera and sensor types to detect flaws, spoilage and other issues. This data is then passed through the rest of the system, where some kind of programming allows for the detection of any issues.

At the board house, suspect PCBs are identified and then taken off the conveyor and handed over to a human who can then either confirm the issue and address or bin it, or mark it as a false positive by the system and put it back on the conveyor. The main advantage here is that it reduces the cognitive load on the humans, who are notoriously terrible at long stretches of boring work.

Another area where MV is essential is that of self-driving vehicles, which is where sensor blending and interpretation of features in a scene using e.g. edge detection and recognition using a convolutional neural network (CNN) is paramount. This replicates the human cognitive process of navigation and steering, though it should be noted that these systems require significant more sensors, including radar and Lidar, to do their job somewhat effectively.

Here it should be noted that MV doesn’t replace human cognition. Rather, it serves to complement it from a general automation perspective. This is why purely self-driving vehicles (Level 5) are still fictional and sometimes comically obvious PCB assembly flaws can make it through automated QA, even if overall it is a net win for the human workers.

Pattern Recognition

Much of the medical profession is about pattern recognition and differential diagnostics, as symptoms and test results have to be categorized and analyzed. Within this field there has been a push towards computer-aided diagnosis (CAD) for decades now, here also to try and reduce the cognitive workload on medical staff. The start of this was with expert systems implemented in e.g. Lisp, which use a knowledge base and an inference system in order to reach a conclusion or solve a problem.

An issue here is of course that this knowledge base has to be constantly maintained, which is why artificial neural network designs have become more popular, with large language models one particular example of these. Such models can be updated more easily, with the slight gotcha that by not having the expert system maintained by human beings any more and instead relying on what are essentially statistical models, you’re abandoning the ‘expert’ part.

This is why LLMs have been increasingly pushed to the side by things like retrieval augmented generation (RAG), which ‘grounds’ the provided facts in more factual reality such as human-written documents, leaving the LLM to help provide a friendly natural language interface.

When it comes to analyzing test results such as of MRI scans and X-rays, this covers much of the same ground as with full MV systems, with the same gotcha that although it can save time, it can also make incredibly dumb mistakes and thus cannot be left unsupervised.

Natural Language

Perhaps the biggest advancement of the past years has been in creating better chatbots that can keep up a conversation on a level that would put ELIZA to shame. Of course, this is at least as much smoke-and-mirrors as ELIZA, in that there is no actual intelligence or concerned therapist behind the friendly interaction, just a complex human-written chat interface that creates the query and handles all other details of using an LLM for generating the semblance of a human-level interaction.

The term ‘emotional intelligence‘ refers to the ability to perceive and feel emotions, something that is impossible for an entity that is incapable of feeling and reasoning, meaning that it is a fairly complex cognitive process that is also heavily susceptible to projection of one’s own feelings onto another person or even an inanimate object. Although the chatbot is literally incapable of learning and requires external session information to be stored within the context window, these can be very convincing near-facsimile under the right conditions.

Faking Cognition

The increased use of machine vision and similar systems has been an absolute boon in automating industries and other fields, making life better for everyone involved due to the reduced cognitive load and freeing up humans to do more creative tasks where one isn’t asked to mindlessly perform the same task over and over.

There are many fields where such increased cognitive offloading is a good thing and quite feasible, but always with a full understanding of the limitations and potential pitfalls, especially when it comes to risks like cognitive atrophy caused by cognitive surrender. This has been identified as a hazard in an increasing number of studies, highlighting the importance of maintaining one’s critical thinking skills.

Even if actual artificial intelligence happened next year, it’s still paramount that we treasure human intelligence, as it is the only one we will always have, as well as the sole reason why humankind has come this far.

from the good-deals-on-cool-stuff dept

Discover the innovative Smart Car Kit, a versatile robot equipped with 4WD omnidirectional movement, FPV (First Person View), app-based remote control, RGB lighting, and a durable metal frame. Powered by an Arduino Uno-compatible main board, this all-in-one kit delivers multi-functional capabilities, including obstacle avoidance, line tracking, IR remote control, face detection, gesture recognition, voice recognition, and C++ programming. Designed for effortless assembly and coding, the kit comes with detailed instructions and tutorials, making it ideal for beginners, hobbyists, educators, and tech enthusiasts aged 15 and above. Dive into robotics and programming with an interactive hands-on experience, supported by comprehensive technical assistance and a vibrant community forum for collaboration and guidance. It’s on sale for $98.

Note: The Techdirt Deals Store is powered and curated by StackSocial. A portion of all sales from Techdirt Deals helps support Techdirt. The products featured do not reflect endorsements by our editorial team.

Filed Under: daily deal

The propensity of gasoline to ‘go stale’ through the process of oxidation is the reason why gasoline that has been stored for a long period of time is considered to be unusable, as it will no longer combust property. Since this process creates the sludge that you find in the bottom of an old gasoline canister, it follows that you may be able to distill out the still good gasoline. With this reasoning, [Joel] over at the [Lowered Expectations] channel set to work to try out this theory.

As part of his job of maintaining things like pressure washers, he got access to many grades of stale gasoline to experiment with. After a short demonstration of how poorly these grades of stale gasoline burn it’s on to the main distillation event. To the stale gasoline aluminium oxide is added as both a catalyst and to create nucleation sites that will prevent ‘bumping’ where you suddenly get a surge out of the heated flask.

Of course, that this is incredibly dangerous should be obvious, and the lack of PPE on the side of [Joel] is somewhat worrying. On the positive side, he does take it easy with ramping up the temperature on the gasoline to try and find the sweet spot where production seems sufficient. This turned out to start at 70°C in the flask when the condenser began to receive its first load of presumably clean-ish gasoline.

The goal here is of course to approximate the function of the fractionating column (‘distillation tower’) at refineries at smaller scale, which [Joel] appears to be doing correctly with what looks to be a Vigreaux column. Since the base product is gasoline with oxidized contaminants this process is of course quite different, so he goes through the different temperature ranges to see what kind of distillate he gets, up to nearly 200°C before calling it.

Ultimately 880 mL of the initial 1 L was collected, with the various distillates combined for testing. Unfortunately none of the testing is actually covered in the video, but it is mentioned at the end that a second batch of the distillate was used to power his car, so presumably it works.

Suffice it to say that ‘works’ doesn’t mean that it is safe, of course. Heating such stale gasoline produces many highly flammable and combustible substances, along with many that are just downright bad for your health to be exposed to. The plethora of very short-term to all the way to very long-term health effects this may cause should be obvious.

In a quick-commerce market obsessed with speed, Indian startup FirstClub has convinced investors that quality may be a fresh opportunity, helping to double its valuation just nine months after its last funding round.

The Bengaluru-based startup has raised $55 million in a Series B round co-led by Peak XV Partners and Sofina, valuing the company at $255 million after the investment. That’s up from $120 million when it last raised capital in September 2025. Existing investors Accel, RTP Global, and Paramark Ventures also participated. The latest financing brings FirstClub’s total funding to $86 million.

As grocery shopping increasingly moves online, India’s quick-commerce market has expanded rapidly, growing from about $6.2 billion in FY25 to an estimated $11 billion-$12 billion in FY26, according to a recent ICICI Securities report. Leading players have popularized online grocery shopping through ever-faster deliveries. However, FirstClub is wagering that a growing segment of consumers will prioritize quality and product curation over receiving orders as quickly as possible.

Founded in 2024 by former Flipkart executive Ayyappan R, FirstClub operates a curated online grocery platform that offers around 4,000 products — roughly a third of the assortment carried by many quick-commerce rivals. The startup says it conducts quality checks on fresh produce, lab-tests certain staples, and works with brands to develop exclusive products, as it seeks to position itself as a trusted destination for groceries rather than a fast-delivery service.

“People don’t need a very large selection, but they need the right quality selection, consistently delivered every single time,” Ayyappan said in an interview.

FirstClub says more than 60% of its customer base consists of women-led households. Unlike many quick-commerce platforms, where staples such as onions, tomatoes, and potatoes dominate sales, Ayyappan said some of FirstClub’s top-selling products include avocados, persimmons, and Modi apples, reflecting demand for premium and curated grocery offerings.

The strategy appears to be resonating with early shoppers. FirstClub says it has crossed 1 million orders and acquired 170,000 households within a year of launching in Bengaluru.

The startup is currently operating at an annualized gross market value (meaning total of all goods sold on its platform) of about $50 million, with customers placing more than four orders a month on average and spending roughly ₹1,200 (about $13) per order, Ayyappan told TechCrunch.

FirstClub plans to use the fresh capital to expand beyond Bengaluru, where it currently operates 21 stores, and deepen its presence in Hyderabad, where it recently launched with three locations. The startup, which employs about 220 people directly, also plans to expand into categories including home and kitchen products, gifting, and other household essentials.

Peak XV Managing Director GV Ravishankar said the firm believes India is seeing the emergence of a larger cohort of affluent, health-conscious consumers willing to pay for higher-quality products, creating space for specialized grocery platforms alongside mainstream quick-commerce players.

“There will be a specific set of consumers who gravitate toward a better-quality platform that serves trustworthy products,” Ravishankar told TechCrunch. “As Indians become wealthier and more informed, there will be more and more people who make that choice.”

Ravishankar compared the trend to the rise of premium grocery chains in developed markets, arguing that India’s retail landscape is beginning to fragment beyond a one-size-fits-all approach centered on price and convenience.

from the j6ers-still-expect-to-get-paid dept

There were some rumors earlier this week that, as it was facing a lot of pushback in court, in the media, and even among (a few) fellow Republicans in Congress, Donald Trump was going to drop his blatantly unconstitutional, illegal, and corrupt $1.776 billion slush fund for MAGA insurrectionists. And now it’s… sorta officially dead… but not really.

Testifying before Congress, Acting Attorney General and full-time Trump toady, told representatives “we are not moving forward with the weaponization fund.”

Acting Attorney General Todd Blanche said Tuesday that the Trump administration is scrapping plans to create a $1.8 billion fund meant to compensate allies of the Republican president after widespread political backlash and setbacks in the courts.

“We are not moving forward with the fund, period,” Blanche said.

But if you watch the actual video, it’s not quite so concrete:

As Rep. Grace Meng asks for Blanche to put it into writing that the fund is not moving forward, Blanche declines to do so. Furthermore, he refuses to say if the equally corrupt, illegal, and problematic deal to not have the IRS audit the president and his family is also going away (meaning that it’s not going away). Of course, he also lies and claims that the settlement is between “the IRS” and Trump, which is simply incorrect. The IRS never signed the agreement. It was done by Blanche and the DOJ. He also tries to (falsely) claim that the audit immunity is “not blanket immunity,” which is just false:

As Rep. DeLauro reads, Blanche’s order (again, not an official “settlement with the IRS despite Blanche’s false claims) gives very clear blanket immunity to Donald Trump, his family, and his businesses:

The United States RELEASES, WAIVES, ACQUITS, and FOREVER DISCHARGES each of the Plaintiffs from, and is hereby FOREVER BARRED and PRECLUDED from prosecuting or pursuing, any and all claims, counterclaims, causes of action, appeals, or requests for any relief, including injunctive relief, monetary relief, damages, examinations or similar or related reviews, appeals, debt relief, costs, attorney’s fees, expenses, and/or interest, whether presently known or unknown, that as of the Effective Date of the Settlement Agreement-have been or could have been asserted by Defendants against any of the Plaintiffs or related or affiliated individuals (including, without limitation, family or others filing jointly), or parties including trusts, parent, sister, or related companies, affiliates, and subsidiaries, by reason of, with respect to, in connection with, or which arise out of (1) any matters that were raised or could have been raised in the Case or the Pending Agency Claims; (2) Lawfare and/or Weaponization; or (3) any matters currently pending or that could be pending (including tax returns filed before the Effective Date) before Defendants or other agencies or departments.

That is blanket immunity. Full stop.

But here’s the tell: if Blanche’s “we are not moving forward, period” were actually true, you’d expect the people who stood to benefit to be disappointed. They’re not. January 6th insurrectionists, including Proud Boys Leader Enrique Tarrio — convicted of seditious conspiracy and later pardoned by Trump — still seem to think they’re going to cash in. Tarrio has been saying he deserves tens of millions of dollars, and rather than expressing any disappointment at Blanche’s testimony, he’s explaining why this is actually good news for him, because it means he can get more money from the US government with less oversight.

If you can’t see that, it’s Tarrio texting reporter Liz Landers:

This isn’t an abandonment. They simply state they’re going to wait two weeks… I believe even if this fund is killed in courts or at a congressional level, the President will find a way… They can just settle the tort claims and lawsuits. That has no judicial review or congressional oversight. And it would mean a lot more money in compensation.

Tarrio’s theory — and likely shared by other J6ers — is that they sue the US government, Trump and Blanche agree to “settle,” and millions of taxpayer dollars flow out through the existing Judgment Fund (the same pot the anti-weaponization fund was drawing from) with zero oversight and zero congressional approval.

And he might not be wrong.

It’s also why a competent Congress would step in and shut all of this down. If we had a competent Congress. Which we don’t.

At some point there needs to be a real reckoning with how broken the system already is — that Trump and Blanche got this far is itself an indictment of how bad things are.

Filed Under: anti-weaponization fund, corruption, donald trump, irs, january 6th, todd blanche

No phishing link. No malware. No SIM swap. Hackers took over high-profile Instagram accounts over the weekend by doing something disarmingly simple: they asked Meta’s AI customer support chatbot to change the email address on someone else’s account. The bot complied without verifying the requester’s identity, and the attacker then reset the password and locked out the rightful owner.

The technique, which was first reported by 404 Media, spread through Telegram channels where hackers shared the method and began advertising stolen handles for sale. Among the compromised accounts were the dormant Obama White House Instagram profile, which was used to post unauthorised AI-generated images, and the account of US Space Force chief master sergeant John Bentivegna.

Meta spokesperson Andy Stone said on Monday that “the issue that did happen has already been fixed.” But on Tuesday, more Instagram users reported losing access to their accounts, and members of the same Telegram channels claimed the exploit still worked, according to TechCrunch.

How the attack worked

The method exploited a flaw in Meta’s AI Support Assistant, which the company rolled out in March 2026 with the ability to “resolve account issues from start to finish,” including resetting passwords. The chatbot was designed to replace human support agents for routine account recovery tasks.

The 💜 of EU tech

Advertisement

The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!

An attacker would identify a target account, typically a short “OG” username worth thousands on underground markets. They would use a VPN to spoof the target’s presumed location, open a chat with the AI support bot, and simply claim to be the account owner. The bot would then link the attacker’s email address to the target account without asking for any proof of ownership.

A human support agent would have verified the caller’s identity before making such a change. The chatbot did not. Two-factor authentication may have blocked some takeovers, but accounts without it enabled were vulnerable to compromise in minutes.

A grey market for stolen handles

For years, a flourishing underground market has existed for so-called OG usernames, the short, desirable handles claimed by Instagram’s earliest users. Previous methods of stealing them required technical sophistication: phishing the victim, bribing telecom insiders to perform SIM swaps, or compromising email accounts.

This attack lowered the barrier to entry dramatically. The hackers who shared the technique on Telegram were advertising apparently stolen handles for sale, including common forenames and country names that function as collectibles in this grey market. TechCrunch reported that the sales continued even after Meta’s announced fix.

Meta scrambles to notify victims

Meta has been sending password reset emails and security notifications to users whose accounts were targeted. Several victims reported receiving messages from Instagram warning that the company had “detected some suspicious activity that suggests your Instagram may have been compromised,” along with instructions to reset their passwords.

Stone told TechCrunch that Meta secured affected accounts on Monday before beginning its notification campaign. He declined to say how many users were compromised. Meta also disputed that the Obama White House account was taken over using this specific method, though it confirmed the account was hacked.

The cost of automating trust

The incident exposes a fundamental tension in deploying AI agents with real-world authority. Meta built its support chatbot to perform actions that previously required a human in the loop, but it shipped that capability without the verification checks that human agents would have applied as a matter of course.

It is a pattern the industry has seen before. When Instagram account recovery was handled by humans, the process was slow and often frustrating, but it at least required the requester to prove they were who they claimed to be. Automating that process without preserving the identity-verification step turned a bottleneck into a vulnerability.

The broader lesson is not that AI should never handle sensitive account operations, but that authentication remains a problem no chatbot can shortcut. Meta gave its AI the power to hand over the keys. The hackers simply walked up and asked for them.