"Real" Entities or Sock Puppets? A New Solution Can Help Analysts and Investigators
January 28, 2025
Bitext’s NAMER (shorthand for "named entity recognition") can deliver precise entity tagging across dozens of languages.
Graphs — knowledge graphs and social graphs — have moved into the mainstream since Leonhard Euler formed the foundation for graph theory in the mid 18th century in Berlin.
With graphs, analysts can take advantage of smart software’s ability to make sense of Named Entity Recognition (NER), event extraction, and relationship mapping.
The problem is that humans change their names (handles, monikers, or aliases) for many reasons: Public embarrassment, a criminal record, a change in marital status, etc.
Bitext’s NER solution, NAMER, is specifically designed to meet the evolving needs of knowledge graph companies, offering exceptional features that tackle industry challenges.
Consider a person disgraced with involvement in a scheme to defraud investors in an artificial intelligence start up. The US Department of Justice published the name of a key actor in this scheme. (Source: https://www.justice.gov/usao-ndca/pr/founder-and-former-ceo-san-francisco-technology-company-and-attorney-indicted-years). The individual was identified by the court as Valerie Lau Beckman. The official court documents used the name "Lau" to reference her involvement in a multi-million dollar scam.
However, in order to correctly identify her in social media, subsequent news stories, and in possible public summaries of her training on a LinkedIn-type of smart software is not enough.
That’s the role of a specialized software solution. Here’s what NAMER delivers.
The system identifies and classifies entities (e.g., people, organizations, locations) in unstructured data. The system accurately links data across different sources of content. The NAMER technology can tag and link significant events (transactions, announcements) to maintain temporal relevance; for example, when Ms. Lau Beckman is discharged from the criminal process. NAMER can connect entities like Ms. Lau or Ms. Beckman to other individuals with whom she works or interacts and her "names" appearance in content streams.
The licensee specifies the languages NAMER is to process, either in a knowledge base or prior to content processing via a large language model.
Access to the proprietary NAMER technology is via a local SDK which is essential for certain types of entity analysis. NAMER can also be integrated into another system or provided as a "white label service" to enhance an intelligence system with NAMER’s unique functions. The developer provides for certain use cases direct access to the source code of the system.
For an organization or investigative team interested in keeping data about Lau Beckman at the highest level of precision, Bitext’s NAMER is an essential service.
Stephen E Arnold, January 28, 2025
China Smart, US Dumb: Some AI Readings in English
January 28, 2025
A blog post from an authentic dinobaby. He’s old; he’s in the sticks; and he is deeply skeptical.
I read a short post in YCombinator’s Hacker News this morning (January 23, 2025). The original article is titled “Deepseek and the Effects of GPU Export Controls.” If you are interested in the poli sci approach to smart software, dive in. However, in the couple of dozen comments on Hacker News to the post, a contributor allegedly named LHL posted some useful links. I have pulled these from the comments and displayed them for your competitive intelligence large language model. On the other hand, you can read them because you are interested in what’s shaking in the Lin-gang Free Trade Zone in the Middle Kingdom:
Deepseek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
Deepseek Coder V2: Breaking the Barrier of Closed Source Models in Code Intelligence
Deepseek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
Deepseek LLM Scaling Open-Source Language Models with Longtermism
First, a thanks to the poster LHL. The search string links timed out, so you may already be part of the HN herd who is looking at the generated bibliography.
Second, several observations:
- China has lots of people. There are numerous highly skilled mathematicians, Monte Carlo and gradient descent wonks, and darned good engineers. One should not assume that wizardry ends with big valuations and tie ups among Oracle, Open AI and the savvy funder of Banjo, an intelware outfit of some repute.
- Computing resource constraints translate into one outcome. Example: Howard Flank, one of my team members, received the Information Industry Association Award decades ago for cramming a searchable index of the Library of Congress’ holdings. Remember those wonderful machines in the early 1980s. Yeah, Howard did wonders with limited resources. The Chinese professionals can too and have. (Note to US government committee members: Keep Howard and similar engineering whiz kids in mind when thinking about how curtailing computer resources will stop innovation.)
- Deepseek’s methods are likely to find there way into some US wrapper products presented as groundbreaking AI. Nope. These innovations are enabled by an open source technology. Now what happens if an outfit like Telegram or one of the many cyber gangs which Microsoft’s Brad Smith references? Yeah. Innovation of a type that is not salubrious.
- The authors of the papers are important. Should these folks be cross correlated with other information about grants, academic affiliations with US institutions, and conference attendance?
In case anyone is curious, from my dinobaby point of view, the most important paper in the bunch is the one about a “mixture of experts.”
Stephen E Arnold, January 28, 2025
Business Strategies That Will Deliver Something, Including Face Time with Legal Eagles
January 28, 2025
A blog post written by a real and still-alive dinobaby. If there is art, there is AI in my workflow.
I read an interesting write up which has an implied forbidden word in its title. I have provided a link, but I made an editorial change to avoid getting into trouble with the oligarchs and their forbidden word punishment squad. The write up is called ““Stealth Mode” and Other … Brilliant Strategies.” Let’s take a look at the list and consider what the “brilliant strategies” mean for a person or entity needing cash, customers, repeat business, a BMW, and probably a yacht or four like Mr. Bezos has.
Here are the strategies, which I would label “tactics”, but dinobabies are persnickety:
- Stealth mode. I think means keep what you are doing a secret.
- Get “it” right. I added the quotes. I am not sure to what the “it” refers.
- Don’t ask anyone if he or she would pay for the product or services.
- Hey, don’t work too hard.
- Write the code first which translates to build the product, software, service, or whatever.
- Raise money before you have customers.
- Copy the competition.
- Sell to everyone.
At first glance, these “strategies” reflect the current business environment. Most of these would be warmly endorsed by the oligarchs who are now helping a new President make big-time decisions. If one assumes that the major technology companies and some of the promising start ups have been using these “brilliant strategies,” how have the knock on effects changed society. Here’s my comment on each of the eight items.
- Stealth mode. I think the idea is to pretend to be doing something in secret. The reality is that information has a tendency to escape. Keeping secrets is difficult. Just ask the major telcos or the victims of a pig butchering scheme. “Stealth mode” is a form of marketing and PR. “Real” journalists love to write about secrets.
- Get it right. Sorry that’s not part of the method today. I would remind you that updates to Microsoft Windows wreck havoc for some people. The security breaches which popular X.com’s OSINT posts are a grim reminder than excellence in software is another marketing precept. The way the world works is, “Ship it Thursday.”
- Don’t ask suggests that market research is not needed. I disagree. One needs input, particularly before creating a product or service. If no one will pay, the organization is dead or dying. Check out the supply of Dodge Hornets. Research may not translate to good decisions.
- Yeah, “don’t work too hard.” Wow. I grew up in a home populated by a father who worked two jobs and a mother who kept house and volunteered for decades at a hospital. The idea of “work” is important. I think that may be why at age 80, I chug along. Working hard has many benefits. These include self respect, money, and a way to remain productive. No, “work hard.” The scrolling life is very bad.
- Build first. Okay, but I think the first two steps are planning and talking to people. Sorry. I stick to the basics.
- Raise money before you have customers. This means that you need to have money. For the top one percenters in income, this may work. For most people, money is needed. Lots of money is needed because shoe string grand slams are infrequent events.
- Copy. Yep, the me-too culture. My view is that one should plan which includes research and analysis. Then figure out how to differentiate and deliver value. Copying Google will get a start up exactly nowhere. Copying content for an AI play will get you some face time with lawyers. Do the Leonardo thing: Be innovative.
- Sell to everyone. Successful people are able to sell their ideas, themselves, narcotics, whatever. The trick here is to talk to people. Those with a need will respond to a solution if it works.,
Let’s step back. Will the budding entrepreneur make a lot of money following these rules? The depressing answer is, “Yes, some carpet baggers will indeed strike it rich.” That’s something to keep in mind. I have some rules of thumb, and they are different from the eight set forth in the write up. Dinobabies are a pain, aren’t they?
Stephen E Arnold, January 28, 2025
Microsoft Still Searching after All These Years
January 28, 2025
Finally, long-suffering Windows users will get a better Windows Search. But only if they are willing to mix AI with their OS. The Register reports, "Improved Windows Search Arrives… But Only for Copilot+ PCs." Reporter Richard Speed writes:
"Windows Search has been the punchline to many a Windows joke over the years. The service is intended to provide an easy way of finding content on a local machine, and has previously been mocked for being slow and unreliable. It was blamed for various failures, from causing high CPU usage and toppling over when bits of infrastructure had issues, to tripping up other applications, such as Outlook. Microsoft is making improvements in the latest Dev Channel release – although only for Copilot+ PCs – and is ‘introducing semantic indexing along with traditional indexing.’ This means typing some natural phrases into the Windows search box on the taskbar or searching in File Explorer, which will produce a list of documents that include items close to or related to the search terms."
For now, Windows Search continues to only work on files stored locally. However, Microsoft plans to expand that to documents in the cloud in a future release. The firm promises "no data gathered during the indexing is sent to the company or used to train AI models." Sure.
How many search systems does Microsoft have? How many work as users expect? Our suggestion: Use Everything search. Not only does that freeware tool work well, it does not require one to embrace AI to function. Oh, Microsoft, keep searching. One day you may find a way to locate information in a Windows system. Maybe?
Cynthia Murrell, January 28, 2025
Rethinking Newspapers: The Dinobaby View
January 27, 2025
A blog post from an authentic dinobaby. He’s old; he’s in the sticks; and he is deeply skeptical.
I read “For Some Newspaper Workers, the New Year Began with Four Weeks of Unpaid Leave.” But the subtitle is the snappy statement:
The chain CNHI furloughed 46 staffers, or about 3% of its workforce. It’s likely a weather vane for industry trouble ahead.
The write up says, rather predictably, in my opinion:
the furloughs were precipitated by a very “soft fourth quarter,” usually the best of the year for newspapers, buoyed with ads for Christmas shopping.
No advertising and Amazon. A one-two punch.
The article concludes:
If you’re looking for a silver lining here, it may be that upstart investors continue to buy up newspapers as they come up for sale, still seeing a potential for profit in the business.
What a newspaper needs is a bit of innovation. Having worked at both newspaper publishing and a magazine publishing companies, I dipped into some of my old lectures about online. I floated these ideas at various times in company talks and in my public lectures, including the one I received from ASIS in the late 1980s. Here’s a selected list:
- People and companies pay for must-have information. Create must-have content in digital form and then sell access to that content.
- Newspapers are intelligence gathering outfits. Focus on intelligence and sell reports to outfits known to purchase these reports.
- Convert to a foundation and get in the grant and fund raising business.
- Online access won’t generate substantial revenue; therefore, use online to promote other information services.
- Each newspaper has a core competency. Convert that core competency into pay-to-attend conferences on specific subjects. Sell booth and exhibit space. Convert selling ads to selling a sponsored cocktail at the event.
- Move from advertising to digital coupons. These can be made available on a simple local-focus Web site. For people who want paper ads, sell a subscription to an envelope containing the coupons and possibly a small amount of information of interest to the area the newspaper serves.
Okay, how many of these ideas are in play today? Most of them, just not from newspaper outfits. That’s the problem. Innovation is tough to spark. Is it too late now? My research team has more ideas. Write benkent2020 at yahoo dot com.
Stephen E Arnold, January 27, 2024
So You Want to Be an AI Millionaire?
January 27, 2025
A blog post from an authentic dinobaby. He’s old; he’s in the sticks; and he is deeply skeptical.
The US Attorney in the Northern District of California issued a remarkable public statement about an AI investor scheme that did not work. The write up states:
A 25-count indictment was unsealed today charging Alexander Charles Beckman, the founder and former CEO of GameOn, Inc., also known as GameOn Technology or ON Platform (“GameOn”), and Valerie Lau Beckman (“Lau”), an attorney who worked on GameOn matters and is married to Beckman, with conspiracy, wire fraud, securities fraud, identity theft, and other offenses. Lau was also charged with obstruction of justice. According to the indictment filed on Jan. 21, 2025, Beckman, 41, and Lau, 38, both of San Francisco, allegedly conspired to defraud GameOn investors, GameOn, and a bank.
I want to point out that this type of fraud is a glimpse of the interesting world of the Silicon Valley FOMO or fear of missing out. Writing checks based on a PowerPoint deck is a variation of playing roulette, just with money not a casino with no clocks.
However, in the official statement, there was some fascinating information about the specific method used by the individuals involved in the scam. The public document says:
As alleged in the indictment, Beckman’s statements to GameOn investors often described non-existent revenue, inflated cash balances, and fake and otherwise exaggerated customer relationships. To further the scheme, Beckman allegedly used the names of at least seven real people—including fake emails and signatures—without their permission to distribute false and fraudulent GameOn financial and business information and documents with the intent to defraud GameOn and its investors. Among the individuals whose names Beckman used to commit the fraud scheme was a GameOn CFO, two bank employees, and an employee of a major professional sports league. Beckman also fabricated two GameOn audit reports using the names, signatures, and trademarks of reputable accounting firms, including one of the Big Four accounting firms, to validate false financial statements, and distributed over a dozen fake bank statements for GameOn’s accounts as part of the scheme.
Building a financial crime is hard, detailed work. Here’s the twist used by those in the US Attorney’s news release:
After changing law firms multiple times, Lau joined a venture capital firm in September 2021. Lau is alleged to have provided Beckman with genuine audit reports that she obtained from her own employer that Beckman then used to create fake audit reports for GameOn. The indictment alleges that Lau personally emailed one of these fake audit reports to a GameOn investor’s representative, knowing it to be fake, to induce further investment into the company. In June 2024, Lau furthered the scheme to defraud by delivering a fake GameOn account statement—one that she knew falsely listed GameOn’s balance at a certain financial institution as over $13 million when the company’s true balance was just $25.93—to a bank branch in San Francisco and asking a bank employee to keep the fake statement in an envelope at the bank for Beckman to pick up later that day. Lau knew that Beckman planned to pick up the fake statement with a GameOn director who represented a major investor on GameOn’s board. Beckman picked up the fake statement with the GameOn director that day.
Several observations:
- Bad actors in this case did a great deal of work. Imagine the benefit of applying those talents to a non-fraudulent activity.
- The FOMO lure generates a pool of suckers for get rich quick schemes.
- The difference between a “real” AI play and one that is little more than a vehicle for big bucks resides on a fine line subject to Heisenberg’s uncertainty principle. Some crazy AI schemes get lucky and become “real” businesses. Everyone is surprised.
The clever work may be rewarded with new career opportunities for those involved.
Stephen E Arnold, January 27, 2025
How to Make Software Smart Like Humans
January 27, 2025
Artificial intelligence algorithms are still only as smart as they’re programmed. In other words, they’re still software, sometimes stupid pieces of software. Most AI algorithms are trained on large language models (LLMS) and datasets that lack the human magic to make them “think” like a smart 14-year-old. That could change says Science Daily based on the research from Linköping University: “Machine Psychology: A Bridge To General AI?”
The Robert Johansson of Linköping University asserted in his dissertation that psychological learning models combined with AI could be the key to making machines smart like humans. Johansson developed the concept of Machine Psychology and explains that, unlike many people, he’s not afraid of an AI future. Artificial General Intelligence (AGI) has many positive and negatives. The technology must be carefully created, but AGI could counter many societal destructive developments.
Johansson suggests that AI developers should follow the principle-led path. He means that through his research he’s identified important psychological learning principles that could explain intelligence and they could be implemented in machines. He’s used a logic system called Non-Axiomatic Reasoning System (NARS) that is purposely designed without complete data, computational power, and in real time. This provides the flexibility to handle problems that arise in reality.
NARS works on limited information like a human:
“The combination of NARS and learning psychology principles constitutes an interdisciplinary approach that Robert Johansson calls Machine Psychology, a concept he was the first to coin but more actors have now started to use, including Google DeepMind. The idea is that artificial intelligence should learn from different experiences during its lifetime and then apply what it has learned to many different situations, just as humans begin to do as early as the age of 18 months — something no other animal can do.”
Johansson said that it is possible machines could be as smart as humans within five years. It is a plan, but do computers have the correct infrastructure to handle that type of intelligence? Do humans have the smarts to handle smarter software?
Whitney Grace, January 27, 2025
Amazon: Twitch Is Looking a Bit Lame
January 24, 2025
Are those 30-second ads driving away viewers? Are the bans working to alienate creators and their fans? Is Amazon going to innovate in streaming?
These are questions Amazon needs to answer in a way that is novel and actually works.
Twitch is an online streaming platform primarily used by gamers to stream their play seasons and interact with their fanbase. There hasn’t been much news about Twitch in recent months and it could be die to declining viewership. Tube Filter dives into the details with “Is Twitch Viewership At Its Lowest Point In Four Years?”
The article explains that Twitch had a total of 1.58 billion watch time hours in December 2024. This was its lowest month in four years according to Stream Charts. Twitch, however, did have a small increase in new streamers joining the platform and the amount of channels live at one time. Stream Charts did mention that December is a slow month due to the holiday season. Twitch is dealing with dire financial straits and made users upset when it used AI to make emotes.
Here are some numbers:
“In both October and November 2024, around 89,000 channels on average would be live on Twitch at any one time. In December, that figure pushed up to 92,392. Twitch also saw a bump in the overall number of active channels from 4,490,725 in November to 4,777,395 in December—a 6% increase. [I]t’s important to note that other key metrics for both viewer and streamer activity remain strong,” it wrote in a report about December’s viewership. “A positive takeaway from December was the variety of content on offer. Streamers broadcasted in 43,200 different categories, the highest figure of the year, second only to March.”
Streams Charts notes that all these streamers broadcasted a more diverse range of content of content than usual.
Twitch is also courting TikTok creators in case the US federal government bans the short video streaming platform. The platform has offerings that streamers want, but it needs to do more to attract more viewers.
Whitney Grace, January 24, 2025
How to Garner Attention from X.com: The Guardian Method Seems Infallible
January 24, 2025
Prepared by a still-alive dinobaby.
The Guardian has revealed its secret to getting social media attention from Twitter (now the X). “‘Just the Start’: X’s New AI Software Driving Online Racist Abuse, Experts Warn” makes the process dead simple. Here are the steps:
- Publish a diatribe about the power of social media in general with specific references to the Twitter machine
- Use name calling to add some clickable bound phrases; for example, “online racism”, “fake images”, and “naked hate”
- Use loaded words to describe images; for example, an athlete “who is black, picking cotton while another shows that same player eating a banana surrounded by monkeys in a forest.”
Bingo. Instantly clickable.
The write up explains:
Callum Hood, the head of research at the Center for Countering Digital Hate (CCDH), said X had become a platform that incentivised and rewarded spreading hate through revenue sharing, and AI imagery made that even easier. “The thing that X has done, to a degree that no other mainstream platform has done, is to offer cash incentives to accounts to do this, so accounts on X are very deliberately posting the most naked hate and disinformation possible.”
This is a recipe for attention and clicks. Will the Guardian be able to convert the magnetism of the method in cash money?
Stephen E Arnold, January 24, 2025
And the Video Game Struggler for 2024 Is… Video Games
January 24, 2025
Yep, 2024 sas the worst year for videogames since 1983.
Videogames are still a young medium, but they’re over fifty years old. The gaming industry has seen ups and downs with the first (and still legendary) being the 1983 crash. Arcade games were all the rage back then, but these days consoles and computers have the action. At least, they should.
Wired writes that “2024 Was The Year The Bottom Fell Out Of The Games Industry” due to multiple reasons. There was massive layoffs in 2023 with over 10,000 game developers losing their jobs. Some of this was attributed to AI slowly replacing developers. The gaming industry’s job loss in 2024 was forty percent higher than the prior year. Yikes!
DEI (diversity, equity, and inclusion) combined with woke mantra was also blamed for the failue of many games, including Suicide Squad: Kill the Justice League. The phrase “go woke, go broke” echoed throughout the industry as it is in Hollywood, Silicon Valley, and other fields. I noted:
“According to Matthew Ball, an adviser and producer in the games and TV space…says that the blame for all of this can’t be pinned to a single thing, like capitalism, mismanagement, Covid-19, or even interest rates. It also involves development costs, how studios are staffed, consumers’ spending habits, and game pricing. “This storm is so brutal,” he says, ‘because it is all of these things at once, and none have really alleviated since the layoffs began.’”
Many indie studios were shuttered and large tech leaders such as Microsoft and Sony shut down parts of their gaming division. Also a chain of events influenced by the hatred of DEI and its associated mindsets that is being called a second GamerGate.
The gaming industry will continue through the beginnings of 2025 with business as usual. The industry will bounce back, but it will be different than the past.
Whitney Grace, January 24, 2025
-
- Subscribe to Beyond Search
Feature archive
News archive
-
