Hacker News Reader

So they racially profiled him by only considering characteristics which excluded race?

Article appears to say the cops started with video of both the murder and a guy at a nearby convenience store that looked like the murderer--according to humans. AI was used to narrow down a short list of people who looked like the guy in the latter video. Cops then decided if any indeed did look like him, and also lived in the right area to appear on those surveillance videos. Not sure if that should be enough, but the use of AI doesn't really seem critical.

> The search turned up what police say is the murder weapon in the suspect’s home.

My understanding is that they connected the weapon to the murder and thus got the criminal. The problem, from my understanding, is that the "connection" was illegal and as a result they have to remove the evidence (the gun). Without that, they have no case.

That's a misunderstanding then. They found a weapon and are claiming a connection, but you should not believe a connection just because the police claim it.

From the article:

> The search turned up what police say is the murder weapon in the suspect’s home.

You can't just pull up a weapon from the suspect home and call it a "murder weapon". They has at least to be some connection especially that everyone and his dog has a weapon in America.

Sure you can. It’s called just making shit up. Police do it all the time, that’s what the article is about.

More importantly, and most people don't realize it, but the police normally find enough evidence to get the suspect to admit and take a plea.

So much of the time they don't need to get evidence to convince a jury, they just need to get the suspect to believe they have enough evidence to convince a jury.

This all "works" because in some large percentage of murder cases, everyone knows who did it, because the "tv show unexplained unexpected murder by someone random of someone they don't even know" is rare.

> You can't just pull up a weapon from the suspect home and call it a "murder weapon".

There is ample history of the police doing this. You should not trust them at face value the same way you do not trust the defendant at face value.

At most I guess they found a cartidge that could have been fired by the suspects weapon. Bullet forensics is not that accurate.

> You can't just pull up a weapon from the suspect home and call it a "murder weapon"

Yes, you can, and without any reference to evidence establishing that it is the murder weapon beyond it being described as “what police say is the murder weapon”, you should absolutely consider the probability that that is exactly what the police are doing.

The way I read it, even if forensic analysis proved that it was the murder weapon, it's inadmissible because the search warrant where it was found shouldn't have been issued in the first place due to Clearview being used to provide the backing for a search warrant.

If the forensic analysis proved that it was the murder weapon, they would leak that detail very clearly to the press, to make the judge look bad.

They couldn’t connect the footage of the murder to anyone, not with AI at the time or now. It is too low quality to ID anyone. They watched live footage 6 days later of the suspect leave his place a block away from the murder location and go to a convenience store. They said he looked like the murderer footage based on “build, hair style, clothing and walking characteristics” which is pretty vague and flimsy. They took images from the convenience store footage and AI face matched it to the suspect, but the AI connection was "illegal" (it isn't illegal to use AI, it just isn't enough for a warrant and they were sloppy and didn't do any more work to get proper justification). The evidence (the gun) was collected illegally and is inadmissible (fruit from a poison tree).

As for whether he did it, and whether the gun is the murder weapon, that is still very much up for debate. They found a gun that they claim is the murder weapon. To actually know that is not easy or simple. My guess is the gun is the same caliber as the one used in the murder and the ammo still in the gun was the same brand/model as the casings or rounds recovered. So given their other evidence leading them to the house, they believe it is the murder weapon. No gunshot residue was found on the gun or clothing (which is not exculpatory, but raises more doubt).

In the end we know that the suspect.

1. Had no physical features drastically different than the murderer (walking with a limp, having an afro, being 5’ 4”, being a woman, etc)

2. Owns a 9mm handgun (standard issue police handguns are 9mm, so its not exactly uncommon)

3. Lives a block away from the murder location.

4. The murderer “stopped” in Tolbert’s apartment complex driveway before the murder.

5. “was seen running toward and away from Tolbert’s apartment immediately after the killing” which I am unsure if that just means they ran down the road and passed by the apartment as he fled.

6. No belongings were found from the victim at Tolbert’s apartment or, as far as I can find, anything else providing any other positive connections.

Story says Clearview returned 8 photos, 2 were Tolbert. Evidence connects Tolbert’s home and driveway to the killer. Why was Clearview part of the warrant at all?

Clearview wasn't mentioned in the affidavit used to get the search warrant, according to the article:

> Legg’s affidavit did not disclose that police used facial recognition as part of their identification.

> The affidavit simply says the fusion center matched Tolbert to “the unidentified male suspect, based on recovered surveillance video.”

(I had to look up what a "fusion center" is, it's a cooperation between different agencies: https://en.wikipedia.org/wiki/Fusion_center)

> The is the biggest indictment of Cleveland PD and the justice system in the entire story - you could choose to call emergency services or you could choose to risk bleeding out because of their reputation. He chose to die.

Oh come on, this sounds absurd. No one is going to commit suicide because the city's EMS has a bad reputation. He was probably in shock and had just had his phone stolen so couldn't call anyone.

>>“build ... walking characteristics,”

>So what I'm getting is... He was a young black man.

Gait analysis is a thing.

>hair style, clothing

If the person had dreads and a hoodey on should they be looking for a guy with short cropped hair wearing a suit?

>Maybe it returned every single black male in the neighborhood and police just picked who they wanted; This is crucially relevant information.

>>The fusion center ran an AI-powered facial recognition search through Clearview AI and emailed its findings to Cleveland police, according to the documents. >>That AI report turned up eight photos, two of which were pictures of Tolbert. *Edit added:>>He also did not indicate that the facial recognition report returned several photos of people other than Tolbert.

Please don't tilt at windmills, the police did a bad thing by not using approved tools and methods to conduct their analysis. You don't need to turn this into something more than what it is.

>>Seriously wounded, Story walked the last mile to his home on School Street. >>His father found him dead in the bathtub the next day.

>The is the biggest indictment of Cleveland PD and the justice system in the entire story - you could choose to call emergency services or you could choose to risk bleeding out because of their reputation. He chose to die.

Or you know, the snitches get stitches culture. Or perhaps once you go into shock and loose too much blood you stop making rational decisions.

>>The search turned up what police say is the murder weapon in the suspect’s home.

>Oh. They said that. No details provided. Oh. Well, that's all right then.

What proof do you need? I understand you don't want to take them at their word but it's not like the police are going to release the full forensic analysis before there is a trial.

In 2020, the Cleveland PD was noted to have 67% of their officers as white while the makeup of the area is 50% black, however it should also be noted that 60% of their leadership; Chief and Deputy Chiefs are black

https://www.wkyc.com/article/news/local/cleveland/a-detailed...

Gait analysis is a thing.

Gait analysis is about as accurate as bite analysis, which the FBI has concluded is less accurate than just flipping a coin. And gait analysis is actually less useful than bite analysis, since a person's gait will change based on circumstances: shoes, clothing, ground, weather, etc.

If the person had dreads and a hoodey on should they be looking for a guy with short cropped hair wearing a suit?

No, the issue is that they only looked at the first guy that their limited facial match search turned up.

What proof do you need? I understand you don't want to take them at their word but it's not like the police are going to release the full forensic analysis before there is a trial.

It's actually quite common for police to release some of the basic forensic details about the crime before the trial, like whether or not the suspect's fingerprints were found on the weapon, especially in cases where the coverage centers around whether they have the right person.