this week, artificial intelligence and human intelligence work together to bring down the crime rate in one california city. advice on preparing for a downturn from an experienced ceo , and dance. that's this week on press:here . of morning, everyone. i'm scott mcgrew, and in every heist movie ever, there's a security guard who was sitting

in front of a bank of video monitors who looks away at just the right moment, and the bad guys -- who of course in the movie are the bad good guys -- get past the camera. >> at midnight, there is a shift change in the security of the buildings. then, we will strike. >> hey, you can count sheep at home. >> why? i get paid to count them here. >> that probably would not happen to security guards at the sentinel. they are human, and they do look away from the monitors from time to time, they are assisted by the cameras themselves, which use artificial intelligence to detect the good guys from the real bad guys and alert the humans. dave salinger is ceo at deep sentinel. thank you for joining

us. i'm intrigued by this, because as we've seen, only humidors monitoring flaws, and so would computers. they would miss things or misinterpret things. but they sure work well together. >> scott, pleasure to be here. that's exactly what we observed . if you have just machines doing something, they are going to fall off the bandwagon at a certain point. we had alexa for many years, and it is still messing up when it's acting on lllook away or worst-case scenario based on their phone playing flappy bird or on instagram. what we saw was, as we analyzed real security in the, the failure ofr own really dominate the world. that's why we have movies that have those scenes were the guards are looking away at just the right moment.

>> and the machines alert the guards, right? so the guards essentially are always dealing with looking at an active situation. it may not be a bad situation, but they are also looking at monitors that machines have alerted them to. >> that's exactly right. what we do to bring them together is, we have the ai and the machine learning responsible for analyzing all the situations and really providing relevant information to the guards. the guards aren't spending time staring at an empty stairwell for 45 minutes. they are looking at events where there something specific in the ai is saying i want you to look at this part of this video because i am concerned about this. to kind of give you an example at my home, the second you step onto my property, as soon as you cross that property line, i have a machine on my home that is analyzing that video in real time. every couple of milliseconds, it is identifying, is there something suspicious happening here?

as soon as it thinks there is something suspicious, it sends it to a lifeguard, and the guard is reviewing the video. the ai has told it here something i want you to look for. if there is something suspicious or actually criminal happening, it triggers the siren and is able to contact the police with a verified call that there somebody doing something criminal, please send an officer immediately. >> there are a lot of subtleties coming up there. let's say it's a storefront i'm trying to protect at midnight. you might be outside my store waiting for huber and my awning is protecting you from the rain. or if it's a home, you could be trying to steal my packages, or you could be the door dash guy. the ai is able to figure out the difference? >> this is one of those places where the ai might be able to say, there might be something happening here. but it's not very good at figuring out the difference between a door dash guy dropping something off and

somebody stealing something. when i looked at videos of actual crimes, a lot of times, criminals will go out of their way. they will wear a construction outfit or a u.p.s. type outfit so they do look like the right person. we've instructed our guards to intervene, with more of a friendly -- a, this is security. i'm responsible for this property and i'm watching you, i am here, and i'm very present. can you tell me what you are doing? >> have you gotten to the point where the ai is learning, or hasn't been set, these are just the parameters you should be doing? >> my background has been in ai for 25 years, and we always wanted to say, it's learning on its own. in the last five years, is a new artificial intelligence called the learning that actually does allow us to do that. so our ai learns not only from general trends, but we learn specific things about specific properties.

we learn in real time now. it really is an amazing technology. and by viewing the tens of thousands of cameras we have all over the united states, we are able to make that learning constant and make it really general whether you are in alaska in the dead of winter, or you are in florida and there's a hurricane going on. we've been able to learn all those different types of scenarios by having this broad set of cameras. >> ai learning is fascinating, because it such a reflection of ourselves. the example i would give you is, i can see in ai figuring out as it looks at all the different types, the times that somebody is trying to break in somewhere. the truth is, men break into places more often than women. that is statistically true. the ai would then tell itself, men are more suspicious than women. which probably is true. yet, it's prejudicial, as well.

>> that's a great question. you know, you're skirting up against some of the even more sensitive areas. >> you can see the trap i'm leading you into. >> what we've done to be really careful about this was to design the artificial intelligence from the get-go to be very aware of potential bias. in fact, we haven't really talked about this very much publicly, but when i started the company, i met with the naacp and the aclu and i ask them for best practices to make sure that we didn't fall into these traps. and then we have been following those guidelines ever since that meeting. and what we found that was probably the most interesting thing was that i was a very first technology executive that had ever reached out to them and asked, how do i prevent these things from happening? they were just blown away to have this conversation at all, and it took about six or seven hours for us to figure out, how

do we reduce this really scary concept to some very specific concepts we can follow that will prevent us from heading down that trap. >> are there things that your system has learned, maybe things you didn't expect them to learn? >> one of the most common things that we have learned is -- is that when we are really hairline trigger, when we identify anything suspicious, the most common problem that we have is flags. and i wouldn't have expected this. but flags, because of the way they wave in the wind, right, if you think about it, we are analyzing 24 frames a second, 80,254 seconds per day. you are talking millions of flames, which kind of looks like someone doing something suspicious outside your door. we've actually had to build a special ai just to identify flags that look like people. and that was definitely not on

our mind when we launched the company. it's one of the key pieces of intellectual property and we patented it. >> that's a movie plot. somebody's going to be watching this interview and then they are going to dress up as a flag. and get past all your -- so i had a smart camera a few years back called lighthouse, and it was really cool, because it was a camera for the inside of your house. but we've set all these parameters, and one of them was, you know, if my dog comes into view, i don't care. if my dog gets up on the kitchen table, which he had this habit of doing, alert me, and i can yell at him. and it was really, really cool. and then lighthouse went out of business because their cameras were expensive. when you are in the hardware business, i get to keep the hardware, but at the software on the back end, which is what you're providing, right? your camera -- it's just a

camera. it's the back and were you are providing the count. have clients raised that concern with you, like, hey, how do i know you're going to be around three years from now if i'm going to make this major investment? >> you ask us frequently. we do provide some continuity for the customers. but the number one thing we do is we do something awesome, quite frankly. when we very first started, we have customers asking us that all the time. and now that we've stocked thousands and thousands -- i think we are at the pace where we are stopping about 20,000 crimes a year. i mean, just massive amounts of crime. and we are a very transparent company. customers see that value. >> you met quite a dent in salina's, i understand. >> that's a great example for us, where we were able to prove that not only can the system stop crime for an individual property, so like my home or my business, but when this technology is applied to a community, it not only stops

the crime for the properties that are protected, but it starts having this kind of halo effect where the attempted crimes for the entire community go down. and what we saw was that initially, when we installed it in some of these businesses, we were stopping three, four, five attempted trespasses and i. that's how bad some of these properties were. after about six months, we saw the attempted trespasses going down by as much as 50, 60, 70%. and so the areas became just less of a soft target and less attractive to potential criminals. >> on an unrelated note, you cofounded the real estate company redfin, and before that, you are working with data over it amazon. this is your first hardware company. as i said, the hardware is a less interesting part. if the software and the ai that still count. were there things about doing hardware that surprised you?

>> yes, yes, and with a little tiny caveat, absolutely. hardware, it's a different ball game in terms of, how can you do this well? how do you integrate with it effectively? we already support other people's cameras and things like that. and as you can probably see behind me and my background, i love hardware. there's a droid that i built that is a fully functional robot that i built by myself, and here are the different prototypes of the cameras we built over the years. so i love that aspect of it, and it's engaging and inciting. but especially with the supply- chain issues that we are seeing during the pandemic and in china right now, it's a real issue. so what i had to do was make sure that as we manage the business and we grew, we started supporting other people's hardware. and that's really what we are focused on right now, is about 70% of our sales now are

actually supporting hardware that we don't build ourselves and is built by a third party. >> all right that makes sense. david selinger is the ceo at deep sentinel. we appreciate you being with us this morning, and press:here will be back here in just a moment.

welcome back to press:here. the latest data on initial jobless claims released on thursday show a slight uptick

in the number of americans asking for unemployment help. now, we don't know that number will get much bigger in the future, but certainly there is pressure on companies, particularly startups, to watch their cash burn. but my next guest says, don't start with the employees when you are trying to cut those costs. his company took a look at how industry spends money on software, and the results were surprising. many paying for software they didn't need. 100% of cfos told you they are planning to spend less money in the organization, and 100% is a pretty alarming number. >> good morning. we were shocked, too, when we saw the numbers saying yes, we have spent quite a bit of money in the last couple of years, and now is the time to tighten the belt. so they are looking to cut costs. >> you look about how -- you

specifically look at what they're spending on software, and a lot of companies overspend. they buy more software licenses than they need. >> thanks to covid, we went remote. and now, we are realizing that we have way too many seats that we actually are using. so these licenses have accumulated in the last four or five years. >> i have the figures in front of me, a design software used for user interface, you found 44% in the company's you pulled had seats that they were using. zoom, 38%. hub spot, marketing and sales software, 30%, and it lassie and , 18%. i think i would be very interested as a cfo if i saw that, wait a minute, 44% of my seat license are being used by

anyone. i just paying a bill. >> if you look at the organization structure, the finance of the last person, to see what you are spending on software, the design team is ahead of everybody. they are building products, and they buy a lot. and then the finance person finds out their are already some that are accumulated. it's definitely shocking for finance to know, but this is what has happened in the growth minded industry we have been in for the last couple of years. >> i worked in a couple of situations where they say log off. if you go too far the other way. but this is sort of your business, right? you help other businesses understand what they are spending? >> that's correct. if you look at how software was bought 20 years ago in larger companies, your five or 10 vendors to choose from. ibm, microsoft, computer associates. today, you and i can put a card

on file and then by a $50 license, and it just gets accumulated. and hundreds of millions of dollars goes to waste when nobody knows what is being spent on what. >> one of the reports that you've offered your clients is something called how to stop being a target of vendor upsizing. once the person from that company has gotten you hooked, he or she comes back and tries to sell me the next thing. >> yeah, and upsizing is a positive term, because your employee numbers are growing, you need more seat. but vendors have become very smart. they have added clauses in contracts which allows them to upsize blindly. so from year one to year two, you will find a clause saying we will automatically upsize you by 3% whether you want it or not. that's a problem we are seeing more and more in the industry. >> your company itself is not immune from possible recession. you said that 100% of the

companies you have surveyed are planning something. what's your strategy? >> we are countercyclical. i think we got lucky with our business model. we help businesses save money. let's say you have $1 million in spending, we can save you 20, 30%, which is $300,000 added to your bottom line. if we do not say be that much, you get 100% of the fees back. so hence, it kind of works win win with us and the cfo in the partnership. >> we seen a lot of venture capital funds also reaching out to portfolio companies, earning them about choppy waters ahead. you are an adviser for sequoia capital. what are you telling other startups that you are advising? >> a simple three things. hey, watch your expense is. software is easy. you can go look at the dashboard, look at the reports,

cut down on the number of seats you have, cut down on your marketing expenses, and any of the other discretionary expense such as food and beverage, traveling -- you can easily dig those things out and have 15, 20% overall savings across the board. and instead of laying off people, you can actually layoff your expenses and save on the staff to keep them on the payroll. >> i wish you the best of luck, the ceo at quolum , thank you for being with us. press:here will be right back.

welcome back to press:here. that was dance church, and online dance class that got a lot of traction during the pandemic as people danced online. some classes had 13,000 students. but the founder of dance church, choreographer kate wollack, has been doing this for a decade, calling it an outlet for the elderly and the uncoordinated. welcome, kate. i have a friend who teaches dance. she does a great job making me feel comfortable to get up there and get out on the dance floor, because i trust her. she's a nice person. but i will often say that wine is also a factor in that situation. how do you get your clients comfortable without the wine?

>> oh, my gosh. i love that. what a great friend to have. well, i like to think that dance is for everyone and everybody is a dancer. and one half of dance church is the music, so all of our classes have such an amazing sort of range of multi-genre pop music. so i like to think that music is one of the most accessible entry points to dance and to the body. so, you know, a room full of amazing, lovely, open people, really great music, and a brilliant dance artist leading an experience for you, that's going to be the thing that gets people moving. >> it is important to set the mood, because going into any situation, whether it's a gym where everyone is lifting weight, or a dance class where people are pretty good at it, and they look really good in this index, getting people comfortable is super important.

>> absolutely. dance church is such an accessible entry point into dance, and to the body. and all of our classes, you don't have to be experienced. you don't have to ever have taken a dance class before. it's all levels, and the teacher leads you and guide you through actions and ordination. it really feels like one part dance party, one part fitness, one part performance. and it really just leaves you to a really fun entrance to an experience. >> obviously, there have been dance classes around for years and jazzercise videos around for years. but you are taking advantage of technology with live streams. in that way, it's a bit like palatine, where people from all over are accessing these dance classes. >> absolutely. the pellets on dance. yeah, you know, the shift during the pandemic was pretty

incredible. like, our community almost demanded dance church to go online, because people use our classes as therapy. they use it for, you know, their mental health, their fitness, and it's also their community. and so when the pandemic shut us down, we were like, this is impossible, this can't happen. we have to solve this problem. and so the fact that my parents can join dance church now whenever they want from the middle of nowhere, in the middle of a cornfield, i think that that is pretty incredible. it has really just become so much more accessible now. >> it's become a phenomenon. what clicked with everyone? as i said, there have been dance classes. there have been jazzercise videos. but something clicked. >> you know, the pandemic was a really scary time. and i think that, you know,

experiencing joy -- experiencing community when you were alone -- that was pretty profound. and i think that, you know, with -- just with dance and sort of popular culture really becoming a thing, dance church has become a really inclusive way for people to actually experience dance without having to learn choreography or learn dances. because all we do in dance church is, we dance. we dance to music however you want to let it out in your body. and so, i think it means something that people turn to joy in their bodies and dance in dark times. and i see that staying. >> now, you have a class and sanford disco. this airs in sunday morning, so this class is less than an hour from now. so you are coming back next

month, your classes also transfer to seattle and portland and chicago -- let me put you on the spot and asked, who were the best dancers? >> like, who are best teachers? >> no, who are your best dancers? chicagoans? >> oh my gosh. i don't know if i can answer that question. everybody is amazing. our communities on the ground in seattle and l.a. and new york, all of these pop up classes, it is just so amazing to see all these people just show up and really just be present for each other. i like to say that dance church is like contagious expression, you know? there is this blurriness as you are in class and sort of getting in your zone, and you see out of the corner of your eye, all of these people dancing , and you kind of catch waves of that and it informs your body. and so every micro community has their own vibe. but it's a very beautiful,

safe, and open space. so anyone who -- i used to say, the only prerequisite is you have to have an open mind. so anyone who can align with that can come to dance church. i don't know, i can't answer your question. >> i like putting you on the spot, and i think that's the exact right answer. kate wallich is the founder of dance church. thank you for joining us here this morning. press:here will be right back.

that's our show for this week. inks to my guest, and think you for making us a part of your sunday morning.

damian trujillo: hello, and welcome to "comunidad del valle." i'm damian trujillo, and today, sabor del valle returns to silicon valley on your "comunidad del valle." ♪♪♪ cc by aberdeen captioning 1-800-688-6621 aberdeen.io damian: we begin today with a program called "supporting our schools." it's about getting our kids ready for back to school. with me on "comunidad del valle" is carlota parra. she is a paraprofessional at alpha blanca alvarado middle school, right here in san jose. carlota, welcome to the show. carlota parra: thank you so much for having me. i'm so excited to share my experience at alpha blanca alvarado.

damian: yeah, thank you for bein' here. well, you know, kids, it seems like they just