1986
During class in second grade we watched the challenger shuttle explode on tv, it was horrifying. This was back when they had a TV on a cart and getting to watch anything was a special event. Everyone was interested in shuttle launches back then and even have a launch on VHS somewhere in our collection, I'm not going to look for it, but I saw it when I was looking for family movies in our old collection. I remember being curious and asking people if they thought that people survived the explosion, then crashed and died. I can't recall where I first heard it, but it stuck with me, I also don't think adults liked these kind of questions. lol
I put so many space shuttle models together as a child, This is something that kids got annually for different events, every store had different models and paints, and things affordable.
1-28-1986
Great Grandpa Harold also (Grandpa Landers stepfather) passed away in January
It was around this time that I started seeing a therapist. This was down on Maple Street, in the same building that I filed a complaint against police officers in the future of 2014
This summer when he went to the camp I could see, I remember my father and the rest of us laying on the dock and looking up at the stars and seeing it I don't remember being amazing because we didn't have a telescope but I remember seeing a streak in the sky, and that inspired me to take pictures with my nephew in 2024 or maybe it was 2025 but now I have the checker
Below is a link to a Youtube video of my May 10, 1986 first communion. You can see us marching along like army troops. I remember taking this event so seriously and now looking back, I feel like it was a complete waste of my life.
You can see my friend Tammy Lynd's and a bunch of other students whose names I've forgotten. lol
During this mass you can see Sister Winnie standing behind me. I'm pretty sure she noticed that she was being recorded. That's why she didn't pinch me on the back of my arms. lol She used to be really mean.
I even got into a scuffle with her, when she tried to pull me down a stairwell once,
and I grabbed the railing, and then she slipped and fell, and then got mad at me because she was being rough with me and I was sick of it.
6:38 The video switches over to us. as kids being rambunctious in my basement on Hope Street. You can see me tossing around this big stuffed monkey, I used to love that thing. It drove me nuts, that stuffed ape had two of the same hands, and by design, his hands were supposed to be able to link them together and it could hang around you neck, but because I had two of the same hands, I couldn't link them together, so he couldn't connect his hands, unless his arms were twisted all funky.
This thing was huge. I thought it was the coolest thing in the world. I used to bring it everywhere.You can see me punching it in this video at 7:03 The video switches over to our porch before it was redone, and you can see my sister as a very small toddler.
That's our cat Durdur on the window sill. 8:12 Video switches over to the family picnic that we held after my first communion. I used to love these family picnics, these gatherings. There's so many people here. My cousins. Michael, Steven, Joey, Kenny, Jeremy and Andrea, my grandpa Stebbins, My grandma stebbins, my uncle Mike, my aunt Sandy, my uncle Dave, my aunt Donna, my aunt Mary. Every time we went over there, it was a blast. We're playing freeze tag in this video. You can hear adults are complaining for me to calm down.
https://youtu.be/3LpO2Ft8zCQ?si=u3sGN9EZKX1VFdc
Summer
We had family picnics, we went over Grandpa Stebbins for Father's Day, we never did that with my mom's side of the family, because they didn't have fathers around. lol
trips to new rok, fern lake
Starting 3rd grade
September
3rd grade was rough. My teacher Mrs. Dubois would pull me in the hallway and ask me what's wrong and say nonsensical things like- I thought you were taking your medication? Why can't you sit still? Why aren't you better?
It was really really stressful and I didn't have the words to explain what I was feeling but there was something really wrong in adults purposely ignored it. At some point I refuse to take the medication because of how bad it was making me feel and a doctor sat with me and alone in an office and threatened to stab me with a needle and said I was going to start getting injections if I didn't take my medication. As an adult I don't handle that kind of stuff very well no ones that have tell me what to do without explaining it, unless you're the Commonwealth of Massachusetts then you can just do it whatever you want lol
⚽ The Hill, the Ball, and the Distance Between Us
(Mary Lynch Field, Springfield, MA – Childhood, 1980s)
You used to love playing soccer—especially when your dad and uncle coached your team. Your brother and cousin were often on the field with you, and your grandmother’s house sat just above the field, making it feel like the whole neighborhood was watching and cheering.
But something changed.
You got sick, and it hit hard.
You said:
“I remember struggling to do leg lifts… I remember struggling to run during games because I didn’t have any energy. There were times I thought soccer sucked, but that’s only cause I was sick.”Unedited Conversations …
Despite that, you still showed up. You ran when you could, you watched when you couldn’t. You wished your dad and uncle would run around with you, just once. And at the end of every practice, there was that ritual:
“We used to run up that hill and touch the trees, then run back down. I do that now with Harley. I ran up as fast as I could—Harley had to catch up. It reminds me of when I was young and I struggled to run up the hill when I was sick. Now I’m so fast.”Unedited Conversations …
You’re not just remembering—you’re rewriting the memory, every time you run it now. And Harley runs with you.
🏥 The Twisted Ankle – Dad, the Station Wagon, and the Rescue
One of the clearest memories you shared:
“One time when we were really young, my dad, my brother and I were playing soccer together at the field at the hill. My father twisted his ankle. I remember panicking and running up the hill to get my mom. She drove our station wagon across the field to pick my dad up.”Unedited Conversations …
It’s one of the first moments you felt the pull of responsibility—not just fear, but action. You ran not just because you were scared, but because you knew someone had to take care of him, and it had to be you.
🛠️ The Hill Was Everything
This field and hill weren’t just for soccer. They were:
A bike ramp and racetrack
A rocket launch site
A place for flying kites, RC cars, and throwing frisbees
Freeze tag, volleyball, log-rolling in summer, and sledding in winter
A GI Joe trench where you joked there might still be a lost ED-209 RoboCop figure buried somewhere
And later, a quiet place where you’d take photos of comets with your nephew
And as you put it:
“Now I’m still showing up on the field—still chasing that ball—even if now it’s with Harley instead of a full team.”
Mary Lynch soccer field
2 adults in back on left My dad Rick on the right, his brother uncle Mark
6 kid in the middle, 1-?, 2-?, 3- Me, 4-?, 5- Bobby, 6- cousin Michael
6 kids in front, 1-?, 2-?, 3- Tommy Cerleli?, 4- cosuin Joey, 5-?. 6-?
🖼️ Image Description: 1980s OLSH Soccer Team Photo – Mary Lynch Field, Springfield, MA
This is a youth soccer team photo taken in the early 1980s on the . The photo includes 12 children and 2 adult coaches. Behind them is the field’s signature hill, scattered with leafless trees that act as directional markers.
Summer
This summer I got to see Halley's comet at aunt Pearl's cabin at Fern Lake in New York. We laid out on the dock and we used binoculars to get a better look at the night sky. I remember being creeped out while we were laying out on the dock, since it was so dark and feeling like it was a little chilly and breezy. I could never shut up during events like this, I was amazed at the fact that there was a streak of something in the sky and I don't think I really understood what it was at the time. I didn't really understand that this is one of those things that you just look at and that's it, I kept expecting something to happen, I almost felt like it was magical.
I would constantly nag my dad to join my brother and I fishing or to go layout on the dock to look at the stars, my mom found me very annoying.
September
I remeber I was having a lot of trouble sitting still at this time, I don't remember specific thought, just restlessness. My most vivid memory of this time is sitting on the stairs outside of the end classroom, on the first floor, I was in the northern stairwell, just staring out the window. I felt like such a disappointment and didn't even understand was expected of me at this point, adults could never explain anything. Mrs. Dubois came out at some point ask why I wasn't better, wasn't I on my medication. I was 8 or 9 years old when she asked this, and I still don't know what she was expecting me to say. lol
I remember spending a lot of time in that stairwell, I remember running up and down the stairs and peeking in the windows, while kids were learning, there are a couple other stairwells that I would run up and down as well, I don't think anyone was ever paying attention to what I was doing. lol
Whenever I would get sent to the office, I would always take the long way I would always end circling around, then backtracking. lol It was always an adventure.
Therapist- Maple St 1986
Down at the First Therapist on Maple Street, I used to love and go playing on these wooden playsets that they had built into the corner. It wasn't anything special, but I thought it was awesome that they built something for kids to climb on inside the waiting room. I didn't know that I was there for mental health help at the time, I didn't understand what these people were trying to get me to talk about. I remember this person showing me pictures of animated characters Anatomy and Physiology and this woman would get me to talk about my knowledge the human body.
Looking back definitely recall that I made woman very uncomfortable with all of my questions, I wanted to know why she was asking me everything, I wanted to what everything did when she pointed at it on the page.
I remember being so curious about why she was asking me these questions. I thought, Why do you want me to know this? What is that for? And when she asked if anybody ever touched me, I remember asking "why would someone want to touch me like that". lol
Almost every visit, we would walk up the street to this candy store and get these little caramel creams, I absolutely love candy, and I think they used this as a way to get me to talk. lol ,
I think this is kind of dumb when talking to someone like me, offering something I like candy, in exchange for a conversation, doesn't really work, I'm willing to talk no matter what, but adding all these distractions got me to think about completely different things. No one ever tried to get me to really focus on how I felt or what I was thinking, and if they did, they didn't pass that information on to my parents, so they were just as clueless as I was.
the needle threat 1986
I remember sitting across the desk from the doctor, I felt so small and intimidated, but the medication that they were giving me at the time made me feel absolutely terrible, and I didn't want to take it anymore. I don't think I was arguing with the Doctor, but at one point he grabbed my right wrist, yanked my arm and squeezed, then threatened to stick me with a needle and start giving me my medication that way, but I was still defiant, I didn't want to take it, it made me feel terrible.
I used to love putting quarters in that stupid gate as we left, The times that my brother would go, we would fight for who got to put the quarter in. lol
October
October: I remember going to Kmart and being able to pick out this G.I. Joe battle platform and I got Main Frame and I'm pretty sure Zandar, I can’t remember what other G.I. Joe’s I got, but I remember seeing Ship Wreck on the shelves and wanting him because he had poly, But I thought main frame was cooler because he was a computer guy.
November
I remember Thanksgiving at being even more exciting than Christmas, We got to go over my grandma Francis/ aunt Flossie's, or Cousin Ellie's, Then we'd also get to go over Grandpa and Grandma Stebbins on Moss Rd> and hang out with all the cousins. Playing GI Joe on the Stonewall in the Stebbins backyard on Moss Road was awesome, there was a little fire pit build into it and we'd play in there too, I never remember having any fires, but I always wanted to.
December
Christmas was awesome, I got the G.I. Joe Mauler, It’s a battery powered tank and you can fit two G.I. Joe’s inside of it and it has a hitch so you could tow other vehicles around. I bought one of these as an adult and had it pull vehicles in the backyard over my father’s when we lived in Warren. Videos as an adult in there somewhere I’ll share them med at some point, or they might be in this blog or somewhere. I know they’re on Facebook and they might be on YouTube
My buddy Andrew Wolan from Putnam shared this with me in September 1, 2026,
Its a video about AI labeled 1985, but appears to be from 1986? Close enough is good for me, I posted this to give people an idea of how we've been lied to about this trash for 40 plus years. Keeping track of data is simple, losing is only good for companies, not families.
In this video
Timeline
Transcript
Search transcript
Search transcript
Chapter 1: Dummy_Label
0:2525 seconds a simple computer runs this elevator system it knows to park the elevator on the ground floor in the morning and on the top floor in the evening based on
0:3333 seconds the pattern of usage but what if everyone here were to suddenly go on the night shift well the elevators would still Park themselves the same way because the computer would require
0:4141 seconds artificial intelligence to automatically adjust to the changing conditions what exactly is artificial intelligence and does it even exist we'll get the answers
0:5050 seconds to these questions as we take a look at Artificial Intelligence on this edition of the computer
0:5858 secondsChronicles
1:171 minute, 17 seconds the computer Chronicles is made possible in part by a grant from apip the American Federation of information processing societies apip sponsors of
1:271 minute, 27 seconds OAC 86 the nation's leading conference on business technology for conference
1:321 minute, 32 secondsinformation call 800 oac1 1986 exploring today's Business Solutions additional
1:391 minute, 39 secondsfunding is provided by McGraw Hill Publishers of bite B's detailed technical articles on new hardware software and languages cover the latest
1:481 minute, 48 secondsin microcomputer technology worldwide welcome to the computer
1:561 minute, 56 secondsChronicles I'm steuart chaet and this is Gary kildall and Gary this is a product called rapor rapor claims to be able to have an intelligent conversation with
2:042 minutes, 4 secondsyou I just asked it if it uses artificial intelligence and it's so dumb it can't even figure out what I'm asking it actually the makers of RoR claim it
2:112 minutes, 11 secondshas AI but they say that's artificial Insanity which may or may not have anything to do with artificial intelligence Gary seriously on AI this is one of those areas in which there
2:192 minutes, 19 secondshave been so many promises but apparently not that much delivery are there really any artificial intelligence products out there today well Stuart the real beauty of artificial intelligence
2:272 minutes, 27 secondsis it's a moving Target once the basic research results in their product it's no longer interesting artificial intelligence work uh 20 years ago Samuels had a checker playing program
2:352 minutes, 35 secondsthat was considered good basic AI research today wouldn't even make it into a Macy's store so we hear about products nowaday nowadays that really have things like uh well they claim
2:442 minutes, 44 secondsthings like natural language processing expert system capability things of that sort we have to now figure out what's the fluff and Dazzle and what's the substance well we're going to try to
2:532 minutes, 53 secondsfind out today by taking a serious look at artificial intelligence we'll meet some of the top names in the country in the AI field and we'll see some fact fascinating demonstrations of artificial
3:013 minutes, 1 secondintelligence at work we're going to begin by taking a look at a vision recognition system that uses AI to interpret what it
3:083 minutes, 8 secondssees human beings are equipped with stereoscopic vision a distinctive feature that allows us to judge distance
3:153 minutes, 15 secondsand the relative placement of objects in addition humans know how to identify the features of a landscape by color shape
3:233 minutes, 23 secondsor placement but how can a computer be programmed to do the same with the same limited information
3:303 minutes, 30 secondsin paloalto California a Machine Vision project is underway to automatically produce digital terrain models by
3:383 minutes, 38 secondsscanning stereoscopic photographs of the Earth's surface working from illumination densities and reflectance
3:463 minutes, 46 secondsvalues the terrain modeling system with some Human Assistance can calculate an image from any point of view it can then
3:533 minutes, 53 secondssynthesize a flight path through the terrain based on any chosen route while the process of creating topographic maps
4:014 minutes, 1 secondis simple geometry for a human being scanning a photograph no matter how detailed is full of
4:084 minutes, 8 secondsambiguities tall objects obscure smaller ones repetitive patterns can be confusing and forms can be
4:174 minutes, 17 secondsmisinterpreted to overcome these ambiguities the SRI project is experimenting with artificial
4:234 minutes, 23 secondsintelligence the goal is to develop a program that can look at the global constraints of an image the parts in
4:314 minutes, 31 secondscontext with the whole like much of today's government funded AI research this program is slanted toward military
4:394 minutes, 39 secondsapplications but its future potential will help us recreate more exotic views from the inside of a human body to the
4:484 minutes, 48 secondsoutside of the Moon joining us now in the studio are Dr
5:005 minutesHarry tenant senior member of the technical staff at Texas Instruments in Austin Texas and Dr s Gerald Kaplan principal technologist for Lotus
5:085 minutes, 8 secondsDevelopment Corporation of Cambridge Massachusetts Gary Jerry the first question um AI is one of those things that uh people are pending to their
5:155 minutes, 15 secondsproducts now to make them fancier add the frosting make them sell better uh it has a chance of uh I guess in a sense being misused and how do you
5:245 minutes, 24 secondscharacterize AI in the in this new technology you're working with well really artificial intelligence is a fundamental software technology that's
5:325 minutes, 32 secondsof Interest primarily to people developing software it's not an enduser product in and of itself that one would want to to sell or to uh there really
5:405 minutes, 40 secondsisn't much of a an AI Market except for the tools market and we're going to see some of the tools here today TR some of these things what are the kinds of
5:475 minutes, 47 secondsthings that you'd say AI can add to well what it can add well first you have to understand what it is it's um really methods for programming using symbols
5:565 minutes, 56 secondsand describing their inter relationships and then being able to reason about about them to reach some kind of conclusions or enhance some existing type of application um what we're going
6:056 minutes, 5 secondsto see in the next few years is a continued growth in the utility and uh intelligence if I can use the word loely programs a little bit easier to use a
6:126 minutes, 12 secondslittle more intelligence more intelligent we're begin we're going to begin to see out of uh Lotus certainly and other companies uh programs that begin to know more about their intended
6:216 minutes, 21 secondsuse and that can help the user by um interpreting the user's intentions as opposed to executing the user's commands
6:286 minutes, 28 secondsOkay okay Harry perfect introduction to what you've got here an example it you've got a natural language system one part of AI first and show us what that
6:356 minutes, 35 secondsis that's right what we have is a way to ask questions of the computer in English so that the user doesn't have to uh
6:436 minutes, 43 secondslearn an Arcane computer language before he gets started in this case what I'm going to be doing is building a sentence up in this portion of the screen out of
6:526 minutes, 52 secondswords and phrases that I select from the menus down below for example the sentence starts with show uh show the
7:007 minutesaverage number of days of sun in and now I'll choose a city in El Paso now I've got a complete sentence and I can ask
7:087 minutes, 8 secondsfor the answer show the average number of days of sun in El Paso 319 days now the benefit of this is that somebody who
7:177 minutes, 17 secondsdoesn't know anything about the kind of database that we're interfacing to or anything about how to command that database can come up and immediately see
7:267 minutes, 26 secondswhat kind of information is in there and can immediately find out how to get the information out Harry can you can you actually do the inverse of this process
7:347 minutes, 34 secondsand type the sentence in immediately or is it something you select you select off the menu and construct the sentence you select off the menu and construct
7:427 minutes, 42 secondsthe sentence I had been doing research in the area of of natural language processing uh which is the the act of
7:507 minutes, 50 secondstyping in the sentences and so on for quite a few years before I came up with this approach the benefit of this approach is that 100% of the query that
7:597 minutes, 59 secondsthe user puts in will be understood because the user finds out what kind of sentences he can put together what kind of things the system compare this with
8:088 minutes, 8 secondswith um products that say where where you actually type sentences in and there's a algorithms and her istics that
8:168 minutes, 16 secondsactually go in there and try and figure out what the sentence is is this the same kind of a product or is this a different kind of thing yeah what we did was if you take away the menus behind
8:258 minutes, 25 secondsthe menus lies this the exact same technology that's in the convention natural language systems uh all we've
8:328 minutes, 32 secondsdone there are limitations to any natural language understanding system and all we've done is to reveal to the user what the extent of that system is
8:418 minutes, 41 secondsand what the limitations are so if if one were to peel this layer off you could actually start typing sentences and it would understand that so the underlying algorithms are like that then
8:498 minutes, 49 secondsthat's right okay okay Harry you have now an expert system you're going to show us and I know you got to reload there if you could do that for a second and Jerry I'd like to ask you now we're
8:578 minutes, 57 secondstrying to get a handle on this definitional problem now what we just saw would you call that artificial intelligence um I would say that that's really in the category of a tool and
9:069 minutes, 6 secondswhat you've seen is an application of that tool to uh creating a database in particular um that's all I would say so who
9:149 minutes, 14 secondsknows I really don't think it's accurate to say one thing is is or is not an AI product AI is not something special that you add to an existing program it's an
9:229 minutes, 22 secondsapproach to programming and creating programs that are flexible or useful or solve problems which previously uh might have required some kind of human intelligence or attention well certainly
9:319 minutes, 31 secondsthe case that there is a gray area there between what we could call a user interface and something that really understands the sense yes and that's I
9:389 minutes, 38 secondsthink I would define AI a little bit differently from what we've heard um AI when it started out in the in the early
9:469 minutes, 46 seconds50s was really a set of problems how can we make computers reason how can we make how can we enable them so we can talk to
9:529 minutes, 52 secondsthem or let them see or other problems like that that we recognize in humans that we do and we associate that with
10:0010 minutesintelligence how can we take that same capability and put it into our artifacts into our computers well it's this range
10:0710 minutes, 7 secondsof problems that make up what is AI we've developed a whole collection of Technologies to address those problems
10:1610 minutes, 16 secondsthe the natural language uh that I've demonstrated is one of the approaches there are many but one of the approaches to that sort of problem show the expert
10:2410 minutes, 24 secondsanother now what we have here is a is a technique for uh decision making and and making of judgments in this particular
10:3210 minutes, 32 secondscase for diagnosing problems with your car uh the the system is asking what what is the problem while my car is
10:4010 minutes, 40 secondsmaking a strange noise so I select uh that choice does the noise change between standing still and driving yes
10:4710 minutes, 47 secondsit does uh does the noise change on a different type of Road well no in this particular case the road doesn't make
10:5410 minutes, 54 secondsany difference is the noise a squeal yes it's a squeal is the noise most obvious when starting the car no for my problem
11:0311 minutes, 3 secondsthat's not the case is the noise most obvious when breaking to a stop yes it is well the system has come to a
11:1011 minutes, 10 secondsdiagnosis of the problem uh and the diagnosis is that the brakes may be glazed the rotors or pads of disc brakes
11:1911 minutes, 19 secondsor the Linings of drum brakes should be inspected for glazing or dirt okay now here in this particular case this is again for clarification uh how does this
11:2711 minutes, 27 secondsdiffer from say just a strict yes or no decision- making process getting you down to the point where that someone has literally put in the all the yes or no
11:3611 minutes, 36 secondsanswers get to the final Point are there judgments and evaluations put on on each one of these uh now the real difference that that comes through expert systems
11:4511 minutes, 45 secondsas opposed to other programming techniques is what Jerry was talking about earlier is the way that the system is programmed in this particular case
11:5311 minutes, 53 secondsthe case of building rule-based expert systems as they're called the the knowledge that goes into it is in the form of a collection of rules the little
12:0212 minutes, 2 secondspieces of information that chained together can build up to make decisions and these rules come from where they they come
12:1112 minutes, 11 secondsfrom an expert somebody who knows how to fix cars now behind the behind this there's is there a a generalized tool that say could be applied to not only
12:1912 minutes, 19 secondsautomobiles but say fixing refrigerators absolutely what's what's behind this is is a tool that can take these rules
12:2612 minutes, 26 secondsthese little pieces of information for whatever context you're in and tie them together to make a a decision or make a
12:3412 minutes, 34 secondsjudgment is this General be applied to to virtually anything did you any any kind of a decision the sense of that if you have an expert that would come and
12:4212 minutes, 42 secondswrite rules about uh literally anything expert right rules about could could this system apply to that the fact is that expert
12:5012 minutes, 50 secondssystems are being applied to a very broad range of applications for example there was one developed at at uh John's
12:5712 minutes, 57 secondsHopkins for diagnosing um uh poisoning situations that you can have on a hotline when somebody calls in
13:0513 minutes, 5 secondsyou find out what to do we've seen that here fact Jerry we have we have just about a minute left you mentioned that at Lotus you are working on some ways to use AI whatever it is in application
13:1313 minutes, 13 secondssoftware give us a clue as to as to what areas where we can see coming up well I think what we're going to see is uh improvements really in three ways the
13:2113 minutes, 21 secondsfirst is new classes of applications and products and of course I'm not about to pre-announce anything so I have to be uh somewhat uh circumspect about talking
13:3013 minutes, 30 secondsabout that um but we're looking into areas of flexible databases uh just certain kinds of decision support systems uh programs that will help you
13:3913 minutes, 39 secondsto keep organized help you to uh carry through rational processes that typically are easier for machines than for people reasoning AIDS we're out of
13:4813 minutes, 48 secondstime I'm I'm awfully sorry now going back to the natural language aspect of artificial intelligence one of the one of the leaders in the country of Stanford University Professor Terry
13:5613 minutes, 56 secondswinegrad and our reporter Wendy Woods went to stf for to talk to professor winegrad professor Terry winegrad of Stanford is today working on two
14:0514 minutes, 5 secondsprojects one programming one political the first is a natural language for programmers which enables them to use
14:1214 minutes, 12 secondsSimple English words referenced to the kind of work the software will be doing the whole Quest has been to move away from having people give very specific
14:2114 minutes, 21 secondsdetails of what the machine should do to putting things more in the terms that make sense to them in their world if I'm doing payroll program I want to think about people and payrolls and money I
14:2814 minutes, 28 secondsdon't want to think about disk files and records and who knows what's in the computer his second project concerns funding for AI research he's worried
14:3814 minutes, 38 secondsthat too many of his colleagues are being lured into lucrative defense department contracts detracting them from basic Research into AI the Star
14:4614 minutes, 46 secondsWars in particular has a very specialized spectrum and I think it's going to distort research to focus much of the money huge amounts of money uh on
14:5514 minutes, 55 secondsthose particular Technologies which are for as I said very fast radiation hard uh Kinds of Kinds of computing currently
15:0215 minutes, 2 secondsup to 50% of computer science research is supported by military grants and only a fraction of computer scientists have
15:1015 minutes, 10 secondsrefused them professor winegrad and dozens of other computer science professors are circulating petitions asking that their colleagues not accept
15:1915 minutes, 19 secondsDepartment of Defense Star Wars funding their belief is that it will drain money from AI projects with practical
15:2615 minutes, 26 secondsapplications increasingly through the of the century for the computer Chronicles I'm Wendy
15:4115 minutes, 41 secondsWoods joining us now in the studio are Gary Hendrick vice president of advanced technology at semantic and next to Gary is Dr ubert dfus a professor at UC
15:5015 minutes, 50 secondsBerkeley the author of what computers can't do the limits of AI and a new book entitled Mind Over machine you really like to start with you and say let's
15:5815 minutes, 58 secondsjust say what what are the kind of fallouts see from AI in terms of of useful products for people well it's not
16:0516 minutes, 5 secondsreally my field I'm a philosopher I'm more interested in the limits of AI we'll start at the bottom
16:1216 minutes, 12 secondsend well the the I don't know about the fallouts I'm on the show to find out about the fallouts so what I want to say is that uh you can't use what's called
16:2016 minutes, 20 secondssymbolic descriptions and rules and inferences to make computers intelligent like we are which doesn't mean that you
16:2716 minutes, 27 secondscan't find many uses for them mhm and uh for that that's not my job what's the reason that you can't do that well the
16:3516 minutes, 35 secondsreason I think is that people Begin by using rules when they're learning some new skill but as they get really good at
16:4216 minutes, 42 secondsit and develop a lot of experience they recognize the situation they're in as similar to one that they've already been in and they just do what worked before
16:5116 minutes, 51 seconds and therefore they don't need to analyze the situation in fact analysis is really for beginners intuition is for experts
16:5916 minutes, 59 seconds so that's where the difference is in intuition that comes into Gary do you disagree with that well I think there's some some truth in it to be sure but um
17:0817 minutes, 8 secondsum we have to start with the premise of what's what's the Enterprise that that we're at now um one Enterprise is the scientific one of trying to push AI to
17:1617 minutes, 16 seconds the limits to find out just exactly what a machine uh is in fact capable of doing another is to take the body of techniques that have been developed so
17:2417 minutes, 24 seconds far and Tred to put those into uh into useful products and um here paired up a philosopher and a very pragmatic person
17:3217 minutes, 32 seconds my own view is that we haven't nearly reached the limits of what we can do and what we need to do is is take those things that we have discovered put them
17:4017 minutes, 40 seconds into machines and make practical tools for uh for people Gary there's a been a lot of activity around Q&A can you show us that product oh I'd be really really
17:4717 minutes, 47 seconds pleased to uh what we have here is the the top uh screen of the intelligent assistant unlike the uh slip through by the way this is called Q&A Q&A that's
17:5617 minutes, 56 seconds right we're taking go ahead um the system that you saw from Texas Instruments a minute ago had you uh make
18:0318 minutes, 3 seconds selections from a menu of uh of choices uh that gave a great deal of control over what could be said uh here we just
18:1018 minutes, 10 seconds let people type in um questions uh pretty much the way that they would like and we uh go through you can see the the
18:1718 minutes, 17 seconds flickering um cursor to analyze them after they've been analyzed we understand what the user wanted to do we
18:2418 minutes, 24 seconds then automatically write a program to uh produce this result by looking at the database and formatting the data and just to make sure that we're in sync
18:3118 minutes, 31 seconds another program looks at that program and and translates it back into English specifically what's it doing so we had here um I want to find the forms for
18:3918 minutes, 39 seconds salespeople who earn more than $155,000 and uh display them for me uh and it translates this into uh select and view
18:4718 minutes, 47 seconds the fors on which the department is sales and the salary is greater than $15,000 and after I see that these two are in fact in agreement I uh press the
18:5518 minutes, 55 seconds button and way we go to the database to to pull up the first uh for which displays information now we can step through these one at a time or zoom in
19:0319 minutes, 3 secondson a whole set of them um having pulled these up off the disc I can go down and pick a particular one and then Zoom back in on on that one as an individual now
19:1119 minutes, 11 seconds the interesting thing about natural language is the ability to uh ask follow-up questions particularly using uh pronouns so um um the next question
19:2119 minutes, 21 secondsthat I'm going to ask here is um now I want to see a report that shows the address and salaries of these people that these people refers back to those
19:2919 minutes, 29 seconds sales folks who are earning more than $15,000 let's let's start for just a second now uh do you feel like an untrained individual can come here and
19:3819 minutes, 38 seconds just type any sentence that Well you certainly can't ask um what's interesting on TV tonight and expect it to uh to work let's say it's it's around
19:4619 minutes, 46 seconds the database I is is there going to be a dialogue going back and forth about well I understand this and please tell me about this or because obviously there's
19:5319 minutes, 53 seconds the there you know if you type the right kinds of sentences they can be parsed properly so what is your feeling what's your EXP so far with that we're certainly not able to parse 100% of the
20:0220 minutes, 2 secondssentences that are typed but if your purpose is not to break the system or to push the limits of AI but rather to get to get a job done and uh you sort of
20:1120 minutes, 11 seconds cooperate with the system then the probability is very high a training process that goes both ways in a got some examp you teach you teach the system and the system teaches you and we
20:1920 minutes, 19 seconds we'll see that in just in just a moment uh maybe we could finish this one just just quickly uh we we asked that these be sorted by state and we also wanted to
20:2820 minutes, 28 seconds see the totals averages it translates that into the internal query uh we go off to the database and and pull the information up and now under the state
20:3620 minutes, 36 seconds of California we get the employees there with their addresses their salaries the total for salary the average for salary and then moving over spreadsheet fashion
20:4320 minutes, 43 seconds we get the same thing for the field that was calculated on the Fly of salary plus bonus and then also spreadsheet style you can walk down through the whole report and finally get the the grand
20:5220 minutes, 52 seconds totals and and the grand averages okay show us the teaching part here Gary all right if we take a uh another sent
21:0021 minutes um that doesn't have a word in it that's known that has a word in it that's not known it'll stop in fact I've asked uh
21:0721 minutes, 7 seconds if an employee has a salary between the average Harvard mans and John's do something and it doesn't know the word man at this point I have a number of
21:1521 minutes, 15 seconds options one of them is to teach the word so I'll select teach I'll select to teach it as a synonym and now I'll say that man is a synonym for a male at
21:2321 minutes, 23 seconds least in this database and having uh done that we um restart the pars process and this new definition has been placed
21:3221 minutes, 32 secondsin the Lexicon so it'll be there next time we we we want to ask a question using the word man uh the question itself is rather interesting um if an
21:3921 minutes, 39 seconds employee has a salary between the average Harvard mans and John's let me see his evaluation and review comments because in an ordinary database management system you'd have to ask
21:4721 minutes, 47 seconds three questions to do this first you'd have to find the salary of the the average Harvard Man then you'd have to find John's salary and finally um do yet
21:5621 minutes, 56 seconds a third pass after writing down these intermediate numbers this does it all for you it also has a problem with John here yes it does because there are two John in the database that tells us about
22:0422 minutes, 4 seconds that now we can U press a button and um it pulls up a little display if we know which John It Is by name we can just select it off if we don't know we can uh
22:1322 minutes, 13 seconds zoom in on the forms for this this one makes 32,000 the other guy makes 100 that's the one that I wanted I'll select uh that John and now we'll make the the
22:2222 minutes, 22 seconds third pass through the database and here's the here's the result based on that important difference between this and just a keyword type of a search you're actually trying to take the sentence apart figure out what the
22:3022 minutes, 30 seconds pieces are it's like just like what you do in about seventh grade and and parsing dri we have just about a minute left now from what you've seen in this
22:3822 minutes, 38 seconds whole program the expert mechanic thing and so on uh have you changed your mind at all I mean aren't they impressive demonstrations of some degree of artificial they are impressive but I
22:4622 minutes, 46 seconds haven't changed my mind because I never thought they couldn't be impressive uh expert systems can do about 80 or 90% as
22:5322 minutes, 53 seconds well as the experts in a field I would have called them competent systems I think they're misnamed but once you are willing to settle for competence which
23:0223 minutes, 2 seconds is important and valuable you can then use rules and inferences to make these confident systems but it's a mistake I
23:1023 minutes, 10 seconds believe to think that if you just work harder you can get to expertise you can get that last 15% then in some areas one doesn't care
23:1923 minutes, 19 seconds I mean if you're going to see a doctor and you're sick you would like to get a doctor that was 100% expert if you could but if you couldn't I'd rather consult
23:2723 minutes, 27 seconds one of the medical experts than my general practitioner because I hear that they're really better another thing if there's time I real briefly okay you can
23:3523 minutes, 35 seconds make there are two kinds of expertise there's a calculating kind and in that area you can even do better than people but most kinds of expertise is of this
23:4223 minutes, 42 seconds intuitive kind and then you settle for less gentlemen thank you we are out of time what are the limits of artificial intelligence and is intelligence uniquely human we asked our commentator
23:5123 minutes, 51 seconds George marrow for his thoughts artificial intelligence such an overly ambitious name for any discipline one
23:5923 minutes, 59 seconds day when we're all more computer literate this expression will take its proper place with other classic contradictions such as Postal Service or
24:0624 minutes, 6 seconds Military Intelligence how could such names such as expert systems or artificial intelligence ever have been coined in
24:1324 minutes, 13 seconds the first place well my theory is that in the late 50s when University based computer research was largely funded by the military these names were invented
24:2224 minutes, 22 seconds to impress the granny agencies as often as not this type of research had little if any connection with reality and more ambitious the sound of the project the
24:3024 minutes, 30 seconds better chance it had to be funded now much of today's AI type software is quite novel and very useful but none of these products will be able to live up
24:3824 minutes, 38 seconds to the image that the term artificial intelligence invokes why not as Dr dfus suggests call an expert system a
24:4524 minutes, 45 seconds competent system on the other hand if a vendor insists that one of his products is intelligent we should insist that the product take a standard IQ test and that
24:5424 minutes, 54 seconds the score be a fixed to the package that's how I see it I'm George Morrow
25:0625 minutes, 6 seconds in the random access file this week it's a new year and projections are already being made as to what kind of a year it will be for computers some industry
25:1425 minutes, 14 seconds Watchers are saying 1986 may be the year that speech recognition finally comes into its own there are three major speech products which could come out
25:2225 minutes, 22 seconds this year first is Ray K's voice writer a word processor that you can dictate to it has has a 10,000-word vocabulary
25:2925 minutes, 29 secondskurile says it just about Works in real time speech systems of Tarzana California is expected to introduce a similar product this year it will have a
25:3825 minutes, 38 seconds 5,000-word vocabulary IBM has a speech recognition project called tangora no word of course and when that may become a product major research projects are
25:4725 minutes, 47 seconds underway at MIT and Carnegie melon to develop talk writers and the Japanese are also investing heavily in speech recognition technology with NEC
25:5525 minutes, 55 seconds Matsushita and Toshiba working on new talk products one major event of the new year will probably be the long awaited
26:0326 minutes, 3 seconds introduction of the IBM lap portable called the clamshell inside word is that there may be two versions of the laptop one with a 3 and 1 half inch drive and
26:1226 minutes, 12 seconds one with no dis Drive the IBM portable with the microf floppy Drive will reportedly come with a utility called x copy that will let you copy any software
26:2026 minutes, 20 seconds from a 5 and A4 in disc drive the significance of IBM moving ahead with a 3 and 1/2 in drive is that it seems likely new IBM desktops will also
26:2926 minutes, 29 secondsconvert to the smaller drives thus setting the stage for The Disappearance of the old 5 and a qu inch format as the standard the other major bit of news due
26:3826 minutes, 38 seconds out of IBM this new year is probably a new graphics version of its top view user interface though some Reports say the new top view 1.2 won't be out until
26:4826 minutes, 48 seconds 1987 Paul Schindler's here now with this week's software review it was for moments like these
26:5526 minutes, 55 seconds that someone invented the automatic telephone dialer what what a pain to look things up the same thing goes for computer Communications there are of
27:0327 minutes, 3 seconds course lots of modems that can dial the phone and lots of software capable of running them now there's a breakthrough product called PFS access you'll
27:1127 minutes, 11 seconds recognize the opening screen and function key assignments although software publishing buys its software from Outsiders it imposes a standard
27:1827 minutes, 18 seconds user interface on them the problem with every other Communications package is setting up log on procedures here no
27:2527 minutes, 25 secondssweat turn on the recording feature log on as you normally would and it takes everything down for you you don't even have to remember to ask PFS access to
27:3427 minutes, 34 seconds save your session it automatically saves every session for regular Communications users like me this is great stuff now
27:4227 minutes, 42 seconds PFS access isn't perfect it lacks some flexibility in uploading files and darn software publishing its copy protected
27:5127 minutes, 51 seconds and that makes it a real pain to use but the gain is worth the pain for PFS access from software publishing in Mountain View California
27:5827 minutes, 58 seconds for the computer Chronicles I'm Paul Schindler well we've heard about it for years but the national computerized Yellow Pages is finally a reality it's
28:0728 minutes, 7 seconds actually called American Business lists and you can access it for 15 bucks a month plus a dollar a minute connect time the company says it's the
28:1528 minutes, 15 seconds equivalent of having a national Yellow Pages directory with 6 million Pages that's it for this week's Chronicles we'll see you next time the computer
28:2328 minutes, 23 seconds Chronicles is made possible in part by a grant from aips the American Federation ation of information processing
28:2928 minutes, 29 seconds societies apip sponsors of OAC 86 the nation's leading conference on business technology for conference information
28:3828 minutes, 38 seconds call 800 oac1 1986 exploring today's Business Solutions additional funding is
28:4528 minutes, 45 seconds provided by McGraw Hill Publishers of bite B's detailed technical articles on new hardware software and languages
28:5328 minutes, 53 seconds cover the latest in microcomputer technology worldwide B the International
29:2329 minutes, 23 secondsstandard
Sync to video time
No comments:
Post a Comment